1 @c Copyright (C) 1988,1989,1992,1993,1994,1996,1998,1999,2000,2001,2002 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
6 @chapter C Implementation-defined behavior
7 @cindex implementation-defined behavior, C language
9 A conforming implementation of ISO C is required to document its
10 choice of behavior in each of the areas that are designated
11 ``implementation defined.'' The following lists all such areas,
12 along with the section number from the ISO/IEC 9899:1999 standard.
15 * Translation implementation::
16 * Environment implementation::
17 * Identifiers implementation::
18 * Characters implementation::
19 * Integers implementation::
20 * Floating point implementation::
21 * Arrays and pointers implementation::
22 * Hints implementation::
23 * Structures unions enumerations and bit-fields implementation::
24 * Qualifiers implementation::
25 * Preprocessing directives implementation::
26 * Library functions implementation::
27 * Architecture implementation::
28 * Locale-specific behavior implementation::
31 @node Translation implementation
36 @cite{How a diagnostic is identified (3.10, 5.1.1.3).}
39 @cite{Whether each nonempty sequence of white-space characters other than
40 new-line is retained or replaced by one space character in translation
44 @node Environment implementation
47 The behavior of these points are dependent on the implementation
48 of the C library, and are not defined by GCC itself.
50 @node Identifiers implementation
55 @cite{Which additional multibyte characters may appear in identifiers
56 and their correspondence to universal character names (6.4.2).}
59 @cite{The number of significant initial characters in an identifier
63 @node Characters implementation
68 @cite{The number of bits in a byte (3.6).}
71 @cite{The values of the members of the execution character set (5.2.1).}
74 @cite{The unique value of the member of the execution character set produced
75 for each of the standard alphabetic escape sequences (5.2.2).}
78 @cite{The value of a @code{char} object into which has been stored any
79 character other than a member of the basic execution character set (6.2.5).}
82 @cite{Which of @code{signed char} or @code{unsigned char} has the same range,
83 representation, and behavior as ``plain'' @code{char} (6.2.5, 6.3.1.1).}
86 @cite{The mapping of members of the source character set (in character
87 constants and string literals) to members of the execution character
88 set (6.4.4.4, 5.1.1.2).}
91 @cite{The value of an integer character constant containing more than one
92 character or containing a character or escape sequence that does not map
93 to a single-byte execution character (6.4.4.4).}
96 @cite{The value of a wide character constant containing more than one
97 multibyte character, or containing a multibyte character or escape
98 sequence not represented in the extended execution character set (6.4.4.4).}
101 @cite{The current locale used to convert a wide character constant consisting
102 of a single multibyte character that maps to a member of the extended
103 execution character set into a corresponding wide character code (6.4.4.4).}
106 @cite{The current locale used to convert a wide string literal into
107 corresponding wide character codes (6.4.5).}
110 @cite{The value of a string literal containing a multibyte character or escape
111 sequence not represented in the execution character set (6.4.5).}
114 @node Integers implementation
119 @cite{Any extended integer types that exist in the implementation (6.2.5).}
122 @cite{Whether signed integer types are represented using sign and magnitude,
123 two's complement, or one's complement, and whether the extraordinary value
124 is a trap representation or an ordinary value (6.2.6.2).}
127 @cite{The rank of any extended integer type relative to another extended
128 integer type with the same precision (6.3.1.1).}
131 @cite{The result of, or the signal raised by, converting an integer to a
132 signed integer type when the value cannot be represented in an object of
133 that type (6.3.1.3).}
136 @cite{The results of some bitwise operations on signed integers (6.5).}
139 @node Floating point implementation
140 @section Floating point
144 @cite{The accuracy of the floating-point operations and of the library
145 functions in @code{<math.h>} and @code{<complex.h>} that return floating-point
146 results (5.2.4.2.2).}
149 @cite{The rounding behaviors characterized by non-standard values
150 of @code{FLT_ROUNDS} @gol
154 @cite{The evaluation methods characterized by non-standard negative
155 values of @code{FLT_EVAL_METHOD} (5.2.4.2.2).}
158 @cite{The direction of rounding when an integer is converted to a
159 floating-point number that cannot exactly represent the original
163 @cite{The direction of rounding when a floating-point number is
164 converted to a narrower floating-point number (6.3.1.5).}
167 @cite{How the nearest representable value or the larger or smaller
168 representable value immediately adjacent to the nearest representable
169 value is chosen for certain floating constants (6.4.4.2).}
172 @cite{Whether and how floating expressions are contracted when not
173 disallowed by the @code{FP_CONTRACT} pragma (6.5).}
176 @cite{The default state for the @code{FENV_ACCESS} pragma (7.6.1).}
179 @cite{Additional floating-point exceptions, rounding modes, environments,
180 and classifications, and their macro names (7.6, 7.12).}
183 @cite{The default state for the @code{FP_CONTRACT} pragma (7.12.2).}
186 @cite{Whether the ``inexact'' floating-point exception can be raised
187 when the rounded result actually does equal the mathematical result
188 in an IEC 60559 conformant implementation (F.9).}
191 @cite{Whether the ``underflow'' (and ``inexact'') floating-point
192 exception can be raised when a result is tiny but not inexact in an
193 IEC 60559 conformant implementation (F.9).}
197 @node Arrays and pointers implementation
198 @section Arrays and pointers
202 @cite{The result of converting a pointer to an integer or
203 vice versa (6.3.2.3).}
205 A cast from pointer to integer discards most-significant bits if the
206 pointer representation is larger than the integer type,
207 sign-extends@footnote{Future versions of GCC may zero-extend, or use
208 a target-defined @code{ptr_extend} pattern. Do not rely on sign extension.}
209 if the pointer representation is smaller than the integer type, otherwise
210 the bits are unchanged.
211 @c ??? We've always claimed that pointers were unsigned entities.
212 @c Shouldn't we therefore be doing zero-extension? If so, the bug
213 @c is in convert_to_integer, where we call type_for_size and request
214 @c a signed integral type. On the other hand, it might be most useful
215 @c for the target if we extend according to POINTERS_EXTEND_UNSIGNED.
217 A cast from integer to pointer discards most-significant bits if the
218 pointer representation is smaller than the integer type, extends according
219 to the signedness of the integer type if the pointer representation
220 is larger than the integer type, otherwise the bits are unchanged.
222 When casting from pointer to integer and back again, the resulting
223 pointer must reference the same object as the original pointer, otherwise
224 the behavior is undefined. That is, one may not use integer arithmetic to
225 avoid the undefined behavior of pointer arithmetic as proscribed in 6.5.6/8.
228 @cite{The size of the result of subtracting two pointers to elements
229 of the same array (6.5.6).}
233 @node Hints implementation
238 @cite{The extent to which suggestions made by using the @code{register}
239 storage-class specifier are effective (6.7.1).}
242 @cite{The extent to which suggestions made by using the inline function
243 specifier are effective (6.7.4).}
247 @node Structures unions enumerations and bit-fields implementation
248 @section Structures, unions, enumerations, and bit-fields
252 @cite{Whether a ``plain'' int bit-field is treated as a @code{signed int}
253 bit-field or as an @code{unsigned int} bit-field (6.7.2, 6.7.2.1).}
256 @cite{Allowable bit-field types other than @code{_Bool}, @code{signed int},
257 and @code{unsigned int} (6.7.2.1).}
260 @cite{Whether a bit-field can straddle a storage-unit boundary (6.7.2.1).}
263 @cite{The order of allocation of bit-fields within a unit (6.7.2.1).}
266 @cite{The alignment of non-bit-field members of structures (6.7.2.1).}
269 @cite{The integer type compatible with each enumerated type (6.7.2.2).}
273 @node Qualifiers implementation
278 @cite{What constitutes an access to an object that has volatile-qualified
283 @node Preprocessing directives implementation
284 @section Preprocessing directives
288 @cite{How sequences in both forms of header names are mapped to headers
289 or external source file names (6.4.7).}
292 @cite{Whether the value of a character constant in a constant expression
293 that controls conditional inclusion matches the value of the same character
294 constant in the execution character set (6.10.1).}
297 @cite{Whether the value of a single-character character constant in a
298 constant expression that controls conditional inclusion may have a
299 negative value (6.10.1).}
302 @cite{The places that are searched for an included @samp{<>} delimited
303 header, and how the places are specified or the header is
304 identified (6.10.2).}
307 @cite{How the named source file is searched for in an included @samp{""}
308 delimited header (6.10.2).}
311 @cite{The method by which preprocessing tokens (possibly resulting from
312 macro expansion) in a @code{#include} directive are combined into a header
316 @cite{The nesting limit for @code{#include} processing (6.10.2).}
319 @cite{Whether the @samp{#} operator inserts a @samp{\} character before
320 the @samp{\} character that begins a universal character name in a
321 character constant or string literal (6.10.3.2).}
324 @cite{The behavior on each recognized non-@code{STDC #pragma}
328 @cite{The definitions for @code{__DATE__} and @code{__TIME__} when
329 respectively, the date and time of translation are not available (6.10.8).}
333 @node Library functions implementation
334 @section Library functions
336 The behavior of these points are dependent on the implementation
337 of the C library, and are not defined by GCC itself.
339 @node Architecture implementation
340 @section Architecture
344 @cite{The values or expressions assigned to the macros specified in the
345 headers @code{<float.h>}, @code{<limits.h>}, and @code{<stdint.h>}
346 (5.2.4.2, 7.18.2, 7.18.3).}
349 @cite{The number, order, and encoding of bytes in any object
350 (when not explicitly specified in this International Standard) (6.2.6.1).}
353 @cite{The value of the result of the sizeof operator (6.5.3.4).}
357 @node Locale-specific behavior implementation
358 @section Locale-specific behavior
360 The behavior of these points are dependent on the implementation
361 of the C library, and are not defined by GCC itself.
364 @chapter Extensions to the C Language Family
365 @cindex extensions, C language
366 @cindex C language extensions
369 GNU C provides several language features not found in ISO standard C@.
370 (The @option{-pedantic} option directs GCC to print a warning message if
371 any of these features is used.) To test for the availability of these
372 features in conditional compilation, check for a predefined macro
373 @code{__GNUC__}, which is always defined under GCC@.
375 These extensions are available in C and Objective-C@. Most of them are
376 also available in C++. @xref{C++ Extensions,,Extensions to the
377 C++ Language}, for extensions that apply @emph{only} to C++.
379 Some features that are in ISO C99 but not C89 or C++ are also, as
380 extensions, accepted by GCC in C89 mode and in C++.
383 * Statement Exprs:: Putting statements and declarations inside expressions.
384 * Local Labels:: Labels local to a statement-expression.
385 * Labels as Values:: Getting pointers to labels, and computed gotos.
386 * Nested Functions:: As in Algol and Pascal, lexical scoping of functions.
387 * Constructing Calls:: Dispatching a call to another function.
388 * Naming Types:: Giving a name to the type of some expression.
389 * Typeof:: @code{typeof}: referring to the type of an expression.
390 * Lvalues:: Using @samp{?:}, @samp{,} and casts in lvalues.
391 * Conditionals:: Omitting the middle operand of a @samp{?:} expression.
392 * Long Long:: Double-word integers---@code{long long int}.
393 * Complex:: Data types for complex numbers.
394 * Hex Floats:: Hexadecimal floating-point constants.
395 * Zero Length:: Zero-length arrays.
396 * Variable Length:: Arrays whose length is computed at run time.
397 * Variadic Macros:: Macros with a variable number of arguments.
398 * Escaped Newlines:: Slightly looser rules for escaped newlines.
399 * Multi-line Strings:: String literals with embedded newlines.
400 * Subscripting:: Any array can be subscripted, even if not an lvalue.
401 * Pointer Arith:: Arithmetic on @code{void}-pointers and function pointers.
402 * Initializers:: Non-constant initializers.
403 * Compound Literals:: Compound literals give structures, unions
405 * Designated Inits:: Labeling elements of initializers.
406 * Cast to Union:: Casting to union type from any member of the union.
407 * Case Ranges:: `case 1 ... 9' and such.
408 * Mixed Declarations:: Mixing declarations and code.
409 * Function Attributes:: Declaring that functions have no side effects,
410 or that they can never return.
411 * Attribute Syntax:: Formal syntax for attributes.
412 * Function Prototypes:: Prototype declarations and old-style definitions.
413 * C++ Comments:: C++ comments are recognized.
414 * Dollar Signs:: Dollar sign is allowed in identifiers.
415 * Character Escapes:: @samp{\e} stands for the character @key{ESC}.
416 * Variable Attributes:: Specifying attributes of variables.
417 * Type Attributes:: Specifying attributes of types.
418 * Alignment:: Inquiring about the alignment of a type or variable.
419 * Inline:: Defining inline functions (as fast as macros).
420 * Extended Asm:: Assembler instructions with C expressions as operands.
421 (With them you can define ``built-in'' functions.)
422 * Constraints:: Constraints for asm operands
423 * Asm Labels:: Specifying the assembler name to use for a C symbol.
424 * Explicit Reg Vars:: Defining variables residing in specified registers.
425 * Alternate Keywords:: @code{__const__}, @code{__asm__}, etc., for header files.
426 * Incomplete Enums:: @code{enum foo;}, with details to follow.
427 * Function Names:: Printable strings which are the name of the current
429 * Return Address:: Getting the return or frame address of a function.
430 * Vector Extensions:: Using vector instructions through built-in functions.
431 * Other Builtins:: Other built-in functions.
432 * Target Builtins:: Built-in functions specific to particular targets.
433 * Pragmas:: Pragmas accepted by GCC.
434 * Unnamed Fields:: Unnamed struct/union fields within structs/unions.
437 @node Statement Exprs
438 @section Statements and Declarations in Expressions
439 @cindex statements inside expressions
440 @cindex declarations inside expressions
441 @cindex expressions containing statements
442 @cindex macros, statements in expressions
444 @c the above section title wrapped and causes an underfull hbox.. i
445 @c changed it from "within" to "in". --mew 4feb93
447 A compound statement enclosed in parentheses may appear as an expression
448 in GNU C@. This allows you to use loops, switches, and local variables
449 within an expression.
451 Recall that a compound statement is a sequence of statements surrounded
452 by braces; in this construct, parentheses go around the braces. For
456 (@{ int y = foo (); int z;
463 is a valid (though slightly more complex than necessary) expression
464 for the absolute value of @code{foo ()}.
466 The last thing in the compound statement should be an expression
467 followed by a semicolon; the value of this subexpression serves as the
468 value of the entire construct. (If you use some other kind of statement
469 last within the braces, the construct has type @code{void}, and thus
470 effectively no value.)
472 This feature is especially useful in making macro definitions ``safe'' (so
473 that they evaluate each operand exactly once). For example, the
474 ``maximum'' function is commonly defined as a macro in standard C as
478 #define max(a,b) ((a) > (b) ? (a) : (b))
482 @cindex side effects, macro argument
483 But this definition computes either @var{a} or @var{b} twice, with bad
484 results if the operand has side effects. In GNU C, if you know the
485 type of the operands (here let's assume @code{int}), you can define
486 the macro safely as follows:
489 #define maxint(a,b) \
490 (@{int _a = (a), _b = (b); _a > _b ? _a : _b; @})
493 Embedded statements are not allowed in constant expressions, such as
494 the value of an enumeration constant, the width of a bit-field, or
495 the initial value of a static variable.
497 If you don't know the type of the operand, you can still do this, but you
498 must use @code{typeof} (@pxref{Typeof}) or type naming (@pxref{Naming
501 Statement expressions are not supported fully in G++, and their fate
502 there is unclear. (It is possible that they will become fully supported
503 at some point, or that they will be deprecated, or that the bugs that
504 are present will continue to exist indefinitely.) Presently, statement
505 expressions do not work well as default arguments.
507 In addition, there are semantic issues with statement-expressions in
508 C++. If you try to use statement-expressions instead of inline
509 functions in C++, you may be surprised at the way object destruction is
510 handled. For example:
513 #define foo(a) (@{int b = (a); b + 3; @})
517 does not work the same way as:
520 inline int foo(int a) @{ int b = a; return b + 3; @}
524 In particular, if the expression passed into @code{foo} involves the
525 creation of temporaries, the destructors for those temporaries will be
526 run earlier in the case of the macro than in the case of the function.
528 These considerations mean that it is probably a bad idea to use
529 statement-expressions of this form in header files that are designed to
530 work with C++. (Note that some versions of the GNU C Library contained
531 header files using statement-expression that lead to precisely this
535 @section Locally Declared Labels
537 @cindex macros, local labels
539 Each statement expression is a scope in which @dfn{local labels} can be
540 declared. A local label is simply an identifier; you can jump to it
541 with an ordinary @code{goto} statement, but only from within the
542 statement expression it belongs to.
544 A local label declaration looks like this:
547 __label__ @var{label};
554 __label__ @var{label1}, @var{label2}, @dots{};
557 Local label declarations must come at the beginning of the statement
558 expression, right after the @samp{(@{}, before any ordinary
561 The label declaration defines the label @emph{name}, but does not define
562 the label itself. You must do this in the usual way, with
563 @code{@var{label}:}, within the statements of the statement expression.
565 The local label feature is useful because statement expressions are
566 often used in macros. If the macro contains nested loops, a @code{goto}
567 can be useful for breaking out of them. However, an ordinary label
568 whose scope is the whole function cannot be used: if the macro can be
569 expanded several times in one function, the label will be multiply
570 defined in that function. A local label avoids this problem. For
574 #define SEARCH(array, target) \
577 typeof (target) _SEARCH_target = (target); \
578 typeof (*(array)) *_SEARCH_array = (array); \
581 for (i = 0; i < max; i++) \
582 for (j = 0; j < max; j++) \
583 if (_SEARCH_array[i][j] == _SEARCH_target) \
584 @{ value = i; goto found; @} \
591 @node Labels as Values
592 @section Labels as Values
593 @cindex labels as values
594 @cindex computed gotos
595 @cindex goto with computed label
596 @cindex address of a label
598 You can get the address of a label defined in the current function
599 (or a containing function) with the unary operator @samp{&&}. The
600 value has type @code{void *}. This value is a constant and can be used
601 wherever a constant of that type is valid. For example:
609 To use these values, you need to be able to jump to one. This is done
610 with the computed goto statement@footnote{The analogous feature in
611 Fortran is called an assigned goto, but that name seems inappropriate in
612 C, where one can do more than simply store label addresses in label
613 variables.}, @code{goto *@var{exp};}. For example,
620 Any expression of type @code{void *} is allowed.
622 One way of using these constants is in initializing a static array that
623 will serve as a jump table:
626 static void *array[] = @{ &&foo, &&bar, &&hack @};
629 Then you can select a label with indexing, like this:
636 Note that this does not check whether the subscript is in bounds---array
637 indexing in C never does that.
639 Such an array of label values serves a purpose much like that of the
640 @code{switch} statement. The @code{switch} statement is cleaner, so
641 use that rather than an array unless the problem does not fit a
642 @code{switch} statement very well.
644 Another use of label values is in an interpreter for threaded code.
645 The labels within the interpreter function can be stored in the
646 threaded code for super-fast dispatching.
648 You may not use this mechanism to jump to code in a different function.
649 If you do that, totally unpredictable things will happen. The best way to
650 avoid this is to store the label address only in automatic variables and
651 never pass it as an argument.
653 An alternate way to write the above example is
656 static const int array[] = @{ &&foo - &&foo, &&bar - &&foo,
658 goto *(&&foo + array[i]);
662 This is more friendly to code living in shared libraries, as it reduces
663 the number of dynamic relocations that are needed, and by consequence,
664 allows the data to be read-only.
666 @node Nested Functions
667 @section Nested Functions
668 @cindex nested functions
669 @cindex downward funargs
672 A @dfn{nested function} is a function defined inside another function.
673 (Nested functions are not supported for GNU C++.) The nested function's
674 name is local to the block where it is defined. For example, here we
675 define a nested function named @code{square}, and call it twice:
679 foo (double a, double b)
681 double square (double z) @{ return z * z; @}
683 return square (a) + square (b);
688 The nested function can access all the variables of the containing
689 function that are visible at the point of its definition. This is
690 called @dfn{lexical scoping}. For example, here we show a nested
691 function which uses an inherited variable named @code{offset}:
695 bar (int *array, int offset, int size)
697 int access (int *array, int index)
698 @{ return array[index + offset]; @}
701 for (i = 0; i < size; i++)
702 @dots{} access (array, i) @dots{}
707 Nested function definitions are permitted within functions in the places
708 where variable definitions are allowed; that is, in any block, before
709 the first statement in the block.
711 It is possible to call the nested function from outside the scope of its
712 name by storing its address or passing the address to another function:
715 hack (int *array, int size)
717 void store (int index, int value)
718 @{ array[index] = value; @}
720 intermediate (store, size);
724 Here, the function @code{intermediate} receives the address of
725 @code{store} as an argument. If @code{intermediate} calls @code{store},
726 the arguments given to @code{store} are used to store into @code{array}.
727 But this technique works only so long as the containing function
728 (@code{hack}, in this example) does not exit.
730 If you try to call the nested function through its address after the
731 containing function has exited, all hell will break loose. If you try
732 to call it after a containing scope level has exited, and if it refers
733 to some of the variables that are no longer in scope, you may be lucky,
734 but it's not wise to take the risk. If, however, the nested function
735 does not refer to anything that has gone out of scope, you should be
738 GCC implements taking the address of a nested function using a technique
739 called @dfn{trampolines}. A paper describing them is available as
742 @uref{http://people.debian.org/~karlheg/Usenix88-lexic.pdf}.
744 A nested function can jump to a label inherited from a containing
745 function, provided the label was explicitly declared in the containing
746 function (@pxref{Local Labels}). Such a jump returns instantly to the
747 containing function, exiting the nested function which did the
748 @code{goto} and any intermediate functions as well. Here is an example:
752 bar (int *array, int offset, int size)
755 int access (int *array, int index)
759 return array[index + offset];
763 for (i = 0; i < size; i++)
764 @dots{} access (array, i) @dots{}
768 /* @r{Control comes here from @code{access}
769 if it detects an error.} */
776 A nested function always has internal linkage. Declaring one with
777 @code{extern} is erroneous. If you need to declare the nested function
778 before its definition, use @code{auto} (which is otherwise meaningless
779 for function declarations).
782 bar (int *array, int offset, int size)
785 auto int access (int *, int);
787 int access (int *array, int index)
791 return array[index + offset];
797 @node Constructing Calls
798 @section Constructing Function Calls
799 @cindex constructing calls
800 @cindex forwarding calls
802 Using the built-in functions described below, you can record
803 the arguments a function received, and call another function
804 with the same arguments, without knowing the number or types
807 You can also record the return value of that function call,
808 and later return that value, without knowing what data type
809 the function tried to return (as long as your caller expects
812 @deftypefn {Built-in Function} {void *} __builtin_apply_args ()
813 This built-in function returns a pointer to data
814 describing how to perform a call with the same arguments as were passed
815 to the current function.
817 The function saves the arg pointer register, structure value address,
818 and all registers that might be used to pass arguments to a function
819 into a block of memory allocated on the stack. Then it returns the
820 address of that block.
823 @deftypefn {Built-in Function} {void *} __builtin_apply (void (*@var{function})(), void *@var{arguments}, size_t @var{size})
824 This built-in function invokes @var{function}
825 with a copy of the parameters described by @var{arguments}
828 The value of @var{arguments} should be the value returned by
829 @code{__builtin_apply_args}. The argument @var{size} specifies the size
830 of the stack argument data, in bytes.
832 This function returns a pointer to data describing
833 how to return whatever value was returned by @var{function}. The data
834 is saved in a block of memory allocated on the stack.
836 It is not always simple to compute the proper value for @var{size}. The
837 value is used by @code{__builtin_apply} to compute the amount of data
838 that should be pushed on the stack and copied from the incoming argument
842 @deftypefn {Built-in Function} {void} __builtin_return (void *@var{result})
843 This built-in function returns the value described by @var{result} from
844 the containing function. You should specify, for @var{result}, a value
845 returned by @code{__builtin_apply}.
849 @section Naming an Expression's Type
852 You can give a name to the type of an expression using a @code{typedef}
853 declaration with an initializer. Here is how to define @var{name} as a
854 type name for the type of @var{exp}:
857 typedef @var{name} = @var{exp};
860 This is useful in conjunction with the statements-within-expressions
861 feature. Here is how the two together can be used to define a safe
862 ``maximum'' macro that operates on any arithmetic type:
866 (@{typedef _ta = (a), _tb = (b); \
867 _ta _a = (a); _tb _b = (b); \
868 _a > _b ? _a : _b; @})
871 @cindex underscores in variables in macros
872 @cindex @samp{_} in variables in macros
873 @cindex local variables in macros
874 @cindex variables, local, in macros
875 @cindex macros, local variables in
877 The reason for using names that start with underscores for the local
878 variables is to avoid conflicts with variable names that occur within the
879 expressions that are substituted for @code{a} and @code{b}. Eventually we
880 hope to design a new form of declaration syntax that allows you to declare
881 variables whose scopes start only after their initializers; this will be a
882 more reliable way to prevent such conflicts.
885 @section Referring to a Type with @code{typeof}
888 @cindex macros, types of arguments
890 Another way to refer to the type of an expression is with @code{typeof}.
891 The syntax of using of this keyword looks like @code{sizeof}, but the
892 construct acts semantically like a type name defined with @code{typedef}.
894 There are two ways of writing the argument to @code{typeof}: with an
895 expression or with a type. Here is an example with an expression:
902 This assumes that @code{x} is an array of pointers to functions;
903 the type described is that of the values of the functions.
905 Here is an example with a typename as the argument:
912 Here the type described is that of pointers to @code{int}.
914 If you are writing a header file that must work when included in ISO C
915 programs, write @code{__typeof__} instead of @code{typeof}.
916 @xref{Alternate Keywords}.
918 A @code{typeof}-construct can be used anywhere a typedef name could be
919 used. For example, you can use it in a declaration, in a cast, or inside
920 of @code{sizeof} or @code{typeof}.
924 This declares @code{y} with the type of what @code{x} points to.
931 This declares @code{y} as an array of such values.
938 This declares @code{y} as an array of pointers to characters:
941 typeof (typeof (char *)[4]) y;
945 It is equivalent to the following traditional C declaration:
951 To see the meaning of the declaration using @code{typeof}, and why it
952 might be a useful way to write, let's rewrite it with these macros:
955 #define pointer(T) typeof(T *)
956 #define array(T, N) typeof(T [N])
960 Now the declaration can be rewritten this way:
963 array (pointer (char), 4) y;
967 Thus, @code{array (pointer (char), 4)} is the type of arrays of 4
968 pointers to @code{char}.
972 @section Generalized Lvalues
973 @cindex compound expressions as lvalues
974 @cindex expressions, compound, as lvalues
975 @cindex conditional expressions as lvalues
976 @cindex expressions, conditional, as lvalues
977 @cindex casts as lvalues
978 @cindex generalized lvalues
979 @cindex lvalues, generalized
980 @cindex extensions, @code{?:}
981 @cindex @code{?:} extensions
982 Compound expressions, conditional expressions and casts are allowed as
983 lvalues provided their operands are lvalues. This means that you can take
984 their addresses or store values into them.
986 Standard C++ allows compound expressions and conditional expressions as
987 lvalues, and permits casts to reference type, so use of this extension
988 is deprecated for C++ code.
990 For example, a compound expression can be assigned, provided the last
991 expression in the sequence is an lvalue. These two expressions are
999 Similarly, the address of the compound expression can be taken. These two
1000 expressions are equivalent:
1007 A conditional expression is a valid lvalue if its type is not void and the
1008 true and false branches are both valid lvalues. For example, these two
1009 expressions are equivalent:
1013 (a ? b = 5 : (c = 5))
1016 A cast is a valid lvalue if its operand is an lvalue. A simple
1017 assignment whose left-hand side is a cast works by converting the
1018 right-hand side first to the specified type, then to the type of the
1019 inner left-hand side expression. After this is stored, the value is
1020 converted back to the specified type to become the value of the
1021 assignment. Thus, if @code{a} has type @code{char *}, the following two
1022 expressions are equivalent:
1026 (int)(a = (char *)(int)5)
1029 An assignment-with-arithmetic operation such as @samp{+=} applied to a cast
1030 performs the arithmetic using the type resulting from the cast, and then
1031 continues as in the previous case. Therefore, these two expressions are
1036 (int)(a = (char *)(int) ((int)a + 5))
1039 You cannot take the address of an lvalue cast, because the use of its
1040 address would not work out coherently. Suppose that @code{&(int)f} were
1041 permitted, where @code{f} has type @code{float}. Then the following
1042 statement would try to store an integer bit-pattern where a floating
1043 point number belongs:
1049 This is quite different from what @code{(int)f = 1} would do---that
1050 would convert 1 to floating point and store it. Rather than cause this
1051 inconsistency, we think it is better to prohibit use of @samp{&} on a cast.
1053 If you really do want an @code{int *} pointer with the address of
1054 @code{f}, you can simply write @code{(int *)&f}.
1057 @section Conditionals with Omitted Operands
1058 @cindex conditional expressions, extensions
1059 @cindex omitted middle-operands
1060 @cindex middle-operands, omitted
1061 @cindex extensions, @code{?:}
1062 @cindex @code{?:} extensions
1064 The middle operand in a conditional expression may be omitted. Then
1065 if the first operand is nonzero, its value is the value of the conditional
1068 Therefore, the expression
1075 has the value of @code{x} if that is nonzero; otherwise, the value of
1078 This example is perfectly equivalent to
1084 @cindex side effect in ?:
1085 @cindex ?: side effect
1087 In this simple case, the ability to omit the middle operand is not
1088 especially useful. When it becomes useful is when the first operand does,
1089 or may (if it is a macro argument), contain a side effect. Then repeating
1090 the operand in the middle would perform the side effect twice. Omitting
1091 the middle operand uses the value already computed without the undesirable
1092 effects of recomputing it.
1095 @section Double-Word Integers
1096 @cindex @code{long long} data types
1097 @cindex double-word arithmetic
1098 @cindex multiprecision arithmetic
1099 @cindex @code{LL} integer suffix
1100 @cindex @code{ULL} integer suffix
1102 ISO C99 supports data types for integers that are at least 64 bits wide,
1103 and as an extension GCC supports them in C89 mode and in C++.
1104 Simply write @code{long long int} for a signed integer, or
1105 @code{unsigned long long int} for an unsigned integer. To make an
1106 integer constant of type @code{long long int}, add the suffix @samp{LL}
1107 to the integer. To make an integer constant of type @code{unsigned long
1108 long int}, add the suffix @samp{ULL} to the integer.
1110 You can use these types in arithmetic like any other integer types.
1111 Addition, subtraction, and bitwise boolean operations on these types
1112 are open-coded on all types of machines. Multiplication is open-coded
1113 if the machine supports fullword-to-doubleword a widening multiply
1114 instruction. Division and shifts are open-coded only on machines that
1115 provide special support. The operations that are not open-coded use
1116 special library routines that come with GCC@.
1118 There may be pitfalls when you use @code{long long} types for function
1119 arguments, unless you declare function prototypes. If a function
1120 expects type @code{int} for its argument, and you pass a value of type
1121 @code{long long int}, confusion will result because the caller and the
1122 subroutine will disagree about the number of bytes for the argument.
1123 Likewise, if the function expects @code{long long int} and you pass
1124 @code{int}. The best way to avoid such problems is to use prototypes.
1127 @section Complex Numbers
1128 @cindex complex numbers
1129 @cindex @code{_Complex} keyword
1130 @cindex @code{__complex__} keyword
1132 ISO C99 supports complex floating data types, and as an extension GCC
1133 supports them in C89 mode and in C++, and supports complex integer data
1134 types which are not part of ISO C99. You can declare complex types
1135 using the keyword @code{_Complex}. As an extension, the older GNU
1136 keyword @code{__complex__} is also supported.
1138 For example, @samp{_Complex double x;} declares @code{x} as a
1139 variable whose real part and imaginary part are both of type
1140 @code{double}. @samp{_Complex short int y;} declares @code{y} to
1141 have real and imaginary parts of type @code{short int}; this is not
1142 likely to be useful, but it shows that the set of complex types is
1145 To write a constant with a complex data type, use the suffix @samp{i} or
1146 @samp{j} (either one; they are equivalent). For example, @code{2.5fi}
1147 has type @code{_Complex float} and @code{3i} has type
1148 @code{_Complex int}. Such a constant always has a pure imaginary
1149 value, but you can form any complex value you like by adding one to a
1150 real constant. This is a GNU extension; if you have an ISO C99
1151 conforming C library (such as GNU libc), and want to construct complex
1152 constants of floating type, you should include @code{<complex.h>} and
1153 use the macros @code{I} or @code{_Complex_I} instead.
1155 @cindex @code{__real__} keyword
1156 @cindex @code{__imag__} keyword
1157 To extract the real part of a complex-valued expression @var{exp}, write
1158 @code{__real__ @var{exp}}. Likewise, use @code{__imag__} to
1159 extract the imaginary part. This is a GNU extension; for values of
1160 floating type, you should use the ISO C99 functions @code{crealf},
1161 @code{creal}, @code{creall}, @code{cimagf}, @code{cimag} and
1162 @code{cimagl}, declared in @code{<complex.h>} and also provided as
1163 built-in functions by GCC@.
1165 @cindex complex conjugation
1166 The operator @samp{~} performs complex conjugation when used on a value
1167 with a complex type. This is a GNU extension; for values of
1168 floating type, you should use the ISO C99 functions @code{conjf},
1169 @code{conj} and @code{conjl}, declared in @code{<complex.h>} and also
1170 provided as built-in functions by GCC@.
1172 GCC can allocate complex automatic variables in a noncontiguous
1173 fashion; it's even possible for the real part to be in a register while
1174 the imaginary part is on the stack (or vice-versa). None of the
1175 supported debugging info formats has a way to represent noncontiguous
1176 allocation like this, so GCC describes a noncontiguous complex
1177 variable as if it were two separate variables of noncomplex type.
1178 If the variable's actual name is @code{foo}, the two fictitious
1179 variables are named @code{foo$real} and @code{foo$imag}. You can
1180 examine and set these two fictitious variables with your debugger.
1182 A future version of GDB will know how to recognize such pairs and treat
1183 them as a single variable with a complex type.
1189 ISO C99 supports floating-point numbers written not only in the usual
1190 decimal notation, such as @code{1.55e1}, but also numbers such as
1191 @code{0x1.fp3} written in hexadecimal format. As a GNU extension, GCC
1192 supports this in C89 mode (except in some cases when strictly
1193 conforming) and in C++. In that format the
1194 @samp{0x} hex introducer and the @samp{p} or @samp{P} exponent field are
1195 mandatory. The exponent is a decimal number that indicates the power of
1196 2 by which the significant part will be multiplied. Thus @samp{0x1.f} is
1203 @samp{p3} multiplies it by 8, and the value of @code{0x1.fp3}
1204 is the same as @code{1.55e1}.
1206 Unlike for floating-point numbers in the decimal notation the exponent
1207 is always required in the hexadecimal notation. Otherwise the compiler
1208 would not be able to resolve the ambiguity of, e.g., @code{0x1.f}. This
1209 could mean @code{1.0f} or @code{1.9375} since @samp{f} is also the
1210 extension for floating-point constants of type @code{float}.
1213 @section Arrays of Length Zero
1214 @cindex arrays of length zero
1215 @cindex zero-length arrays
1216 @cindex length-zero arrays
1217 @cindex flexible array members
1219 Zero-length arrays are allowed in GNU C@. They are very useful as the
1220 last element of a structure which is really a header for a variable-length
1229 struct line *thisline = (struct line *)
1230 malloc (sizeof (struct line) + this_length);
1231 thisline->length = this_length;
1234 In ISO C89, you would have to give @code{contents} a length of 1, which
1235 means either you waste space or complicate the argument to @code{malloc}.
1237 In ISO C99, you would use a @dfn{flexible array member}, which is
1238 slightly different in syntax and semantics:
1242 Flexible array members are written as @code{contents[]} without
1246 Flexible array members have incomplete type, and so the @code{sizeof}
1247 operator may not be applied. As a quirk of the original implementation
1248 of zero-length arrays, @code{sizeof} evaluates to zero.
1251 Flexible array members may only appear as the last member of a
1252 @code{struct} that is otherwise non-empty.
1255 GCC versions before 3.0 allowed zero-length arrays to be statically
1256 initialized, as if they were flexible arrays. In addition to those
1257 cases that were useful, it also allowed initializations in situations
1258 that would corrupt later data. Non-empty initialization of zero-length
1259 arrays is now treated like any case where there are more initializer
1260 elements than the array holds, in that a suitable warning about "excess
1261 elements in array" is given, and the excess elements (all of them, in
1262 this case) are ignored.
1264 Instead GCC allows static initialization of flexible array members.
1265 This is equivalent to defining a new structure containing the original
1266 structure followed by an array of sufficient size to contain the data.
1267 I.e.@: in the following, @code{f1} is constructed as if it were declared
1273 @} f1 = @{ 1, @{ 2, 3, 4 @} @};
1276 struct f1 f1; int data[3];
1277 @} f2 = @{ @{ 1 @}, @{ 2, 3, 4 @} @};
1281 The convenience of this extension is that @code{f1} has the desired
1282 type, eliminating the need to consistently refer to @code{f2.f1}.
1284 This has symmetry with normal static arrays, in that an array of
1285 unknown size is also written with @code{[]}.
1287 Of course, this extension only makes sense if the extra data comes at
1288 the end of a top-level object, as otherwise we would be overwriting
1289 data at subsequent offsets. To avoid undue complication and confusion
1290 with initialization of deeply nested arrays, we simply disallow any
1291 non-empty initialization except when the structure is the top-level
1292 object. For example:
1295 struct foo @{ int x; int y[]; @};
1296 struct bar @{ struct foo z; @};
1298 struct foo a = @{ 1, @{ 2, 3, 4 @} @}; // @r{Valid.}
1299 struct bar b = @{ @{ 1, @{ 2, 3, 4 @} @} @}; // @r{Invalid.}
1300 struct bar c = @{ @{ 1, @{ @} @} @}; // @r{Valid.}
1301 struct foo d[1] = @{ @{ 1 @{ 2, 3, 4 @} @} @}; // @r{Invalid.}
1304 @node Variable Length
1305 @section Arrays of Variable Length
1306 @cindex variable-length arrays
1307 @cindex arrays of variable length
1310 Variable-length automatic arrays are allowed in ISO C99, and as an
1311 extension GCC accepts them in C89 mode and in C++. (However, GCC's
1312 implementation of variable-length arrays does not yet conform in detail
1313 to the ISO C99 standard.) These arrays are
1314 declared like any other automatic arrays, but with a length that is not
1315 a constant expression. The storage is allocated at the point of
1316 declaration and deallocated when the brace-level is exited. For
1321 concat_fopen (char *s1, char *s2, char *mode)
1323 char str[strlen (s1) + strlen (s2) + 1];
1326 return fopen (str, mode);
1330 @cindex scope of a variable length array
1331 @cindex variable-length array scope
1332 @cindex deallocating variable length arrays
1333 Jumping or breaking out of the scope of the array name deallocates the
1334 storage. Jumping into the scope is not allowed; you get an error
1337 @cindex @code{alloca} vs variable-length arrays
1338 You can use the function @code{alloca} to get an effect much like
1339 variable-length arrays. The function @code{alloca} is available in
1340 many other C implementations (but not in all). On the other hand,
1341 variable-length arrays are more elegant.
1343 There are other differences between these two methods. Space allocated
1344 with @code{alloca} exists until the containing @emph{function} returns.
1345 The space for a variable-length array is deallocated as soon as the array
1346 name's scope ends. (If you use both variable-length arrays and
1347 @code{alloca} in the same function, deallocation of a variable-length array
1348 will also deallocate anything more recently allocated with @code{alloca}.)
1350 You can also use variable-length arrays as arguments to functions:
1354 tester (int len, char data[len][len])
1360 The length of an array is computed once when the storage is allocated
1361 and is remembered for the scope of the array in case you access it with
1364 If you want to pass the array first and the length afterward, you can
1365 use a forward declaration in the parameter list---another GNU extension.
1369 tester (int len; char data[len][len], int len)
1375 @cindex parameter forward declaration
1376 The @samp{int len} before the semicolon is a @dfn{parameter forward
1377 declaration}, and it serves the purpose of making the name @code{len}
1378 known when the declaration of @code{data} is parsed.
1380 You can write any number of such parameter forward declarations in the
1381 parameter list. They can be separated by commas or semicolons, but the
1382 last one must end with a semicolon, which is followed by the ``real''
1383 parameter declarations. Each forward declaration must match a ``real''
1384 declaration in parameter name and data type. ISO C99 does not support
1385 parameter forward declarations.
1387 @node Variadic Macros
1388 @section Macros with a Variable Number of Arguments.
1389 @cindex variable number of arguments
1390 @cindex macro with variable arguments
1391 @cindex rest argument (in macro)
1392 @cindex variadic macros
1394 In the ISO C standard of 1999, a macro can be declared to accept a
1395 variable number of arguments much as a function can. The syntax for
1396 defining the macro is similar to that of a function. Here is an
1400 #define debug(format, ...) fprintf (stderr, format, __VA_ARGS__)
1403 Here @samp{@dots{}} is a @dfn{variable argument}. In the invocation of
1404 such a macro, it represents the zero or more tokens until the closing
1405 parenthesis that ends the invocation, including any commas. This set of
1406 tokens replaces the identifier @code{__VA_ARGS__} in the macro body
1407 wherever it appears. See the CPP manual for more information.
1409 GCC has long supported variadic macros, and used a different syntax that
1410 allowed you to give a name to the variable arguments just like any other
1411 argument. Here is an example:
1414 #define debug(format, args...) fprintf (stderr, format, args)
1417 This is in all ways equivalent to the ISO C example above, but arguably
1418 more readable and descriptive.
1420 GNU CPP has two further variadic macro extensions, and permits them to
1421 be used with either of the above forms of macro definition.
1423 In standard C, you are not allowed to leave the variable argument out
1424 entirely; but you are allowed to pass an empty argument. For example,
1425 this invocation is invalid in ISO C, because there is no comma after
1432 GNU CPP permits you to completely omit the variable arguments in this
1433 way. In the above examples, the compiler would complain, though since
1434 the expansion of the macro still has the extra comma after the format
1437 To help solve this problem, CPP behaves specially for variable arguments
1438 used with the token paste operator, @samp{##}. If instead you write
1441 #define debug(format, ...) fprintf (stderr, format, ## __VA_ARGS__)
1444 and if the variable arguments are omitted or empty, the @samp{##}
1445 operator causes the preprocessor to remove the comma before it. If you
1446 do provide some variable arguments in your macro invocation, GNU CPP
1447 does not complain about the paste operation and instead places the
1448 variable arguments after the comma. Just like any other pasted macro
1449 argument, these arguments are not macro expanded.
1451 @node Escaped Newlines
1452 @section Slightly Looser Rules for Escaped Newlines
1453 @cindex escaped newlines
1454 @cindex newlines (escaped)
1456 Recently, the preprocessor has relaxed its treatment of escaped
1457 newlines. Previously, the newline had to immediately follow a
1458 backslash. The current implementation allows whitespace in the form of
1459 spaces, horizontal and vertical tabs, and form feeds between the
1460 backslash and the subsequent newline. The preprocessor issues a
1461 warning, but treats it as a valid escaped newline and combines the two
1462 lines to form a single logical line. This works within comments and
1463 tokens, including multi-line strings, as well as between tokens.
1464 Comments are @emph{not} treated as whitespace for the purposes of this
1465 relaxation, since they have not yet been replaced with spaces.
1467 @node Multi-line Strings
1468 @section String Literals with Embedded Newlines
1469 @cindex multi-line string literals
1471 As an extension, GNU CPP permits string literals to cross multiple lines
1472 without escaping the embedded newlines. Each embedded newline is
1473 replaced with a single @samp{\n} character in the resulting string
1474 literal, regardless of what form the newline took originally.
1476 CPP currently allows such strings in directives as well (other than the
1477 @samp{#include} family). This is deprecated and will eventually be
1481 @section Non-Lvalue Arrays May Have Subscripts
1482 @cindex subscripting
1483 @cindex arrays, non-lvalue
1485 @cindex subscripting and function values
1486 In ISO C99, arrays that are not lvalues still decay to pointers, and
1487 may be subscripted, although they may not be modified or used after
1488 the next sequence point and the unary @samp{&} operator may not be
1489 applied to them. As an extension, GCC allows such arrays to be
1490 subscripted in C89 mode, though otherwise they do not decay to
1491 pointers outside C99 mode. For example,
1492 this is valid in GNU C though not valid in C89:
1496 struct foo @{int a[4];@};
1502 return f().a[index];
1508 @section Arithmetic on @code{void}- and Function-Pointers
1509 @cindex void pointers, arithmetic
1510 @cindex void, size of pointer to
1511 @cindex function pointers, arithmetic
1512 @cindex function, size of pointer to
1514 In GNU C, addition and subtraction operations are supported on pointers to
1515 @code{void} and on pointers to functions. This is done by treating the
1516 size of a @code{void} or of a function as 1.
1518 A consequence of this is that @code{sizeof} is also allowed on @code{void}
1519 and on function types, and returns 1.
1521 @opindex Wpointer-arith
1522 The option @option{-Wpointer-arith} requests a warning if these extensions
1526 @section Non-Constant Initializers
1527 @cindex initializers, non-constant
1528 @cindex non-constant initializers
1530 As in standard C++ and ISO C99, the elements of an aggregate initializer for an
1531 automatic variable are not required to be constant expressions in GNU C@.
1532 Here is an example of an initializer with run-time varying elements:
1535 foo (float f, float g)
1537 float beat_freqs[2] = @{ f-g, f+g @};
1542 @node Compound Literals
1543 @section Compound Literals
1544 @cindex constructor expressions
1545 @cindex initializations in expressions
1546 @cindex structures, constructor expression
1547 @cindex expressions, constructor
1548 @cindex compound literals
1549 @c The GNU C name for what C99 calls compound literals was "constructor expressions".
1551 ISO C99 supports compound literals. A compound literal looks like
1552 a cast containing an initializer. Its value is an object of the
1553 type specified in the cast, containing the elements specified in
1554 the initializer; it is an lvalue. As an extension, GCC supports
1555 compound literals in C89 mode and in C++.
1557 Usually, the specified type is a structure. Assume that
1558 @code{struct foo} and @code{structure} are declared as shown:
1561 struct foo @{int a; char b[2];@} structure;
1565 Here is an example of constructing a @code{struct foo} with a compound literal:
1568 structure = ((struct foo) @{x + y, 'a', 0@});
1572 This is equivalent to writing the following:
1576 struct foo temp = @{x + y, 'a', 0@};
1581 You can also construct an array. If all the elements of the compound literal
1582 are (made up of) simple constant expressions, suitable for use in
1583 initializers of objects of static storage duration, then the compound
1584 literal can be coerced to a pointer to its first element and used in
1585 such an initializer, as shown here:
1588 char **foo = (char *[]) @{ "x", "y", "z" @};
1591 Compound literals for scalar types and union types are is
1592 also allowed, but then the compound literal is equivalent
1595 As a GNU extension, GCC allows initialization of objects with static storage
1596 duration by compound literals (which is not possible in ISO C99, because
1597 the initializer is not a constant).
1598 It is handled as if the object was initialized only with the bracket
1599 enclosed list if compound literal's and object types match.
1600 The initializer list of the compound literal must be constant.
1601 If the object being initialized has array type of unknown size, the size is
1602 determined by compound literal size.
1605 static struct foo x = (struct foo) @{1, 'a', 'b'@};
1606 static int y[] = (int []) @{1, 2, 3@};
1607 static int z[] = (int [3]) @{1@};
1611 The above lines are equivalent to the following:
1613 static struct foo x = @{1, 'a', 'b'@};
1614 static int y[] = @{1, 2, 3@};
1615 static int z[] = @{1, 0, 0@};
1618 @node Designated Inits
1619 @section Designated Initializers
1620 @cindex initializers with labeled elements
1621 @cindex labeled elements in initializers
1622 @cindex case labels in initializers
1623 @cindex designated initializers
1625 Standard C89 requires the elements of an initializer to appear in a fixed
1626 order, the same as the order of the elements in the array or structure
1629 In ISO C99 you can give the elements in any order, specifying the array
1630 indices or structure field names they apply to, and GNU C allows this as
1631 an extension in C89 mode as well. This extension is not
1632 implemented in GNU C++.
1634 To specify an array index, write
1635 @samp{[@var{index}] =} before the element value. For example,
1638 int a[6] = @{ [4] = 29, [2] = 15 @};
1645 int a[6] = @{ 0, 0, 15, 0, 29, 0 @};
1649 The index values must be constant expressions, even if the array being
1650 initialized is automatic.
1652 An alternative syntax for this which has been obsolete since GCC 2.5 but
1653 GCC still accepts is to write @samp{[@var{index}]} before the element
1654 value, with no @samp{=}.
1656 To initialize a range of elements to the same value, write
1657 @samp{[@var{first} ... @var{last}] = @var{value}}. This is a GNU
1658 extension. For example,
1661 int widths[] = @{ [0 ... 9] = 1, [10 ... 99] = 2, [100] = 3 @};
1665 If the value in it has side-effects, the side-effects will happen only once,
1666 not for each initialized field by the range initializer.
1669 Note that the length of the array is the highest value specified
1672 In a structure initializer, specify the name of a field to initialize
1673 with @samp{.@var{fieldname} =} before the element value. For example,
1674 given the following structure,
1677 struct point @{ int x, y; @};
1681 the following initialization
1684 struct point p = @{ .y = yvalue, .x = xvalue @};
1691 struct point p = @{ xvalue, yvalue @};
1694 Another syntax which has the same meaning, obsolete since GCC 2.5, is
1695 @samp{@var{fieldname}:}, as shown here:
1698 struct point p = @{ y: yvalue, x: xvalue @};
1702 The @samp{[@var{index}]} or @samp{.@var{fieldname}} is known as a
1703 @dfn{designator}. You can also use a designator (or the obsolete colon
1704 syntax) when initializing a union, to specify which element of the union
1705 should be used. For example,
1708 union foo @{ int i; double d; @};
1710 union foo f = @{ .d = 4 @};
1714 will convert 4 to a @code{double} to store it in the union using
1715 the second element. By contrast, casting 4 to type @code{union foo}
1716 would store it into the union as the integer @code{i}, since it is
1717 an integer. (@xref{Cast to Union}.)
1719 You can combine this technique of naming elements with ordinary C
1720 initialization of successive elements. Each initializer element that
1721 does not have a designator applies to the next consecutive element of the
1722 array or structure. For example,
1725 int a[6] = @{ [1] = v1, v2, [4] = v4 @};
1732 int a[6] = @{ 0, v1, v2, 0, v4, 0 @};
1735 Labeling the elements of an array initializer is especially useful
1736 when the indices are characters or belong to an @code{enum} type.
1741 = @{ [' '] = 1, ['\t'] = 1, ['\h'] = 1,
1742 ['\f'] = 1, ['\n'] = 1, ['\r'] = 1 @};
1745 @cindex designator lists
1746 You can also write a series of @samp{.@var{fieldname}} and
1747 @samp{[@var{index}]} designators before an @samp{=} to specify a
1748 nested subobject to initialize; the list is taken relative to the
1749 subobject corresponding to the closest surrounding brace pair. For
1750 example, with the @samp{struct point} declaration above:
1753 struct point ptarray[10] = @{ [2].y = yv2, [2].x = xv2, [0].x = xv0 @};
1757 If the same field is initialized multiple times, it will have value from
1758 the last initialization. If any such overridden initialization has
1759 side-effect, it is unspecified whether the side-effect happens or not.
1760 Currently, gcc will discard them and issue a warning.
1763 @section Case Ranges
1765 @cindex ranges in case statements
1767 You can specify a range of consecutive values in a single @code{case} label,
1771 case @var{low} ... @var{high}:
1775 This has the same effect as the proper number of individual @code{case}
1776 labels, one for each integer value from @var{low} to @var{high}, inclusive.
1778 This feature is especially useful for ranges of ASCII character codes:
1784 @strong{Be careful:} Write spaces around the @code{...}, for otherwise
1785 it may be parsed wrong when you use it with integer values. For example,
1800 @section Cast to a Union Type
1801 @cindex cast to a union
1802 @cindex union, casting to a
1804 A cast to union type is similar to other casts, except that the type
1805 specified is a union type. You can specify the type either with
1806 @code{union @var{tag}} or with a typedef name. A cast to union is actually
1807 a constructor though, not a cast, and hence does not yield an lvalue like
1808 normal casts. (@xref{Compound Literals}.)
1810 The types that may be cast to the union type are those of the members
1811 of the union. Thus, given the following union and variables:
1814 union foo @{ int i; double d; @};
1820 both @code{x} and @code{y} can be cast to type @code{union foo}.
1822 Using the cast as the right-hand side of an assignment to a variable of
1823 union type is equivalent to storing in a member of the union:
1828 u = (union foo) x @equiv{} u.i = x
1829 u = (union foo) y @equiv{} u.d = y
1832 You can also use the union cast as a function argument:
1835 void hack (union foo);
1837 hack ((union foo) x);
1840 @node Mixed Declarations
1841 @section Mixed Declarations and Code
1842 @cindex mixed declarations and code
1843 @cindex declarations, mixed with code
1844 @cindex code, mixed with declarations
1846 ISO C99 and ISO C++ allow declarations and code to be freely mixed
1847 within compound statements. As an extension, GCC also allows this in
1848 C89 mode. For example, you could do:
1857 Each identifier is visible from where it is declared until the end of
1858 the enclosing block.
1860 @node Function Attributes
1861 @section Declaring Attributes of Functions
1862 @cindex function attributes
1863 @cindex declaring attributes of functions
1864 @cindex functions that never return
1865 @cindex functions that have no side effects
1866 @cindex functions in arbitrary sections
1867 @cindex functions that behave like malloc
1868 @cindex @code{volatile} applied to function
1869 @cindex @code{const} applied to function
1870 @cindex functions with @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style arguments
1871 @cindex functions that are passed arguments in registers on the 386
1872 @cindex functions that pop the argument stack on the 386
1873 @cindex functions that do not pop the argument stack on the 386
1875 In GNU C, you declare certain things about functions called in your program
1876 which help the compiler optimize function calls and check your code more
1879 The keyword @code{__attribute__} allows you to specify special
1880 attributes when making a declaration. This keyword is followed by an
1881 attribute specification inside double parentheses. The following
1882 attributes are currently defined for functions on all targets:
1883 @code{noreturn}, @code{noinline}, @code{always_inline},
1884 @code{pure}, @code{const},
1885 @code{format}, @code{format_arg}, @code{no_instrument_function},
1886 @code{section}, @code{constructor}, @code{destructor}, @code{used},
1887 @code{unused}, @code{deprecated}, @code{weak}, @code{malloc}, and
1888 @code{alias}. Several other attributes are defined for functions on
1889 particular target systems. Other attributes, including @code{section}
1890 are supported for variables declarations (@pxref{Variable Attributes})
1891 and for types (@pxref{Type Attributes}).
1893 You may also specify attributes with @samp{__} preceding and following
1894 each keyword. This allows you to use them in header files without
1895 being concerned about a possible macro of the same name. For example,
1896 you may use @code{__noreturn__} instead of @code{noreturn}.
1898 @xref{Attribute Syntax}, for details of the exact syntax for using
1902 @cindex @code{noreturn} function attribute
1904 A few standard library functions, such as @code{abort} and @code{exit},
1905 cannot return. GCC knows this automatically. Some programs define
1906 their own functions that never return. You can declare them
1907 @code{noreturn} to tell the compiler this fact. For example,
1911 void fatal () __attribute__ ((noreturn));
1916 @dots{} /* @r{Print error message.} */ @dots{}
1922 The @code{noreturn} keyword tells the compiler to assume that
1923 @code{fatal} cannot return. It can then optimize without regard to what
1924 would happen if @code{fatal} ever did return. This makes slightly
1925 better code. More importantly, it helps avoid spurious warnings of
1926 uninitialized variables.
1928 Do not assume that registers saved by the calling function are
1929 restored before calling the @code{noreturn} function.
1931 It does not make sense for a @code{noreturn} function to have a return
1932 type other than @code{void}.
1934 The attribute @code{noreturn} is not implemented in GCC versions
1935 earlier than 2.5. An alternative way to declare that a function does
1936 not return, which works in the current version and in some older
1937 versions, is as follows:
1940 typedef void voidfn ();
1942 volatile voidfn fatal;
1945 @cindex @code{noinline} function attribute
1947 This function attribute prevents a function from being considered for
1950 @cindex @code{always_inline} function attribute
1952 Generally, functions are not inlined unless optimization is specified.
1953 For functions declared inline, this attribute inlines the function even
1954 if no optimization level was specified.
1956 @cindex @code{pure} function attribute
1958 Many functions have no effects except the return value and their
1959 return value depends only on the parameters and/or global variables.
1960 Such a function can be subject
1961 to common subexpression elimination and loop optimization just as an
1962 arithmetic operator would be. These functions should be declared
1963 with the attribute @code{pure}. For example,
1966 int square (int) __attribute__ ((pure));
1970 says that the hypothetical function @code{square} is safe to call
1971 fewer times than the program says.
1973 Some of common examples of pure functions are @code{strlen} or @code{memcmp}.
1974 Interesting non-pure functions are functions with infinite loops or those
1975 depending on volatile memory or other system resource, that may change between
1976 two consecutive calls (such as @code{feof} in a multithreading environment).
1978 The attribute @code{pure} is not implemented in GCC versions earlier
1980 @cindex @code{const} function attribute
1982 Many functions do not examine any values except their arguments, and
1983 have no effects except the return value. Basically this is just slightly
1984 more strict class than the @code{pure} attribute above, since function is not
1985 allowed to read global memory.
1987 @cindex pointer arguments
1988 Note that a function that has pointer arguments and examines the data
1989 pointed to must @emph{not} be declared @code{const}. Likewise, a
1990 function that calls a non-@code{const} function usually must not be
1991 @code{const}. It does not make sense for a @code{const} function to
1994 The attribute @code{const} is not implemented in GCC versions earlier
1995 than 2.5. An alternative way to declare that a function has no side
1996 effects, which works in the current version and in some older versions,
2000 typedef int intfn ();
2002 extern const intfn square;
2005 This approach does not work in GNU C++ from 2.6.0 on, since the language
2006 specifies that the @samp{const} must be attached to the return value.
2009 @item format (@var{archetype}, @var{string-index}, @var{first-to-check})
2010 @cindex @code{format} function attribute
2012 The @code{format} attribute specifies that a function takes @code{printf},
2013 @code{scanf}, @code{strftime} or @code{strfmon} style arguments which
2014 should be type-checked against a format string. For example, the
2019 my_printf (void *my_object, const char *my_format, ...)
2020 __attribute__ ((format (printf, 2, 3)));
2024 causes the compiler to check the arguments in calls to @code{my_printf}
2025 for consistency with the @code{printf} style format string argument
2028 The parameter @var{archetype} determines how the format string is
2029 interpreted, and should be @code{printf}, @code{scanf}, @code{strftime}
2030 or @code{strfmon}. (You can also use @code{__printf__},
2031 @code{__scanf__}, @code{__strftime__} or @code{__strfmon__}.) The
2032 parameter @var{string-index} specifies which argument is the format
2033 string argument (starting from 1), while @var{first-to-check} is the
2034 number of the first argument to check against the format string. For
2035 functions where the arguments are not available to be checked (such as
2036 @code{vprintf}), specify the third parameter as zero. In this case the
2037 compiler only checks the format string for consistency. For
2038 @code{strftime} formats, the third parameter is required to be zero.
2040 In the example above, the format string (@code{my_format}) is the second
2041 argument of the function @code{my_print}, and the arguments to check
2042 start with the third argument, so the correct parameters for the format
2043 attribute are 2 and 3.
2045 @opindex ffreestanding
2046 The @code{format} attribute allows you to identify your own functions
2047 which take format strings as arguments, so that GCC can check the
2048 calls to these functions for errors. The compiler always (unless
2049 @option{-ffreestanding} is used) checks formats
2050 for the standard library functions @code{printf}, @code{fprintf},
2051 @code{sprintf}, @code{scanf}, @code{fscanf}, @code{sscanf}, @code{strftime},
2052 @code{vprintf}, @code{vfprintf} and @code{vsprintf} whenever such
2053 warnings are requested (using @option{-Wformat}), so there is no need to
2054 modify the header file @file{stdio.h}. In C99 mode, the functions
2055 @code{snprintf}, @code{vsnprintf}, @code{vscanf}, @code{vfscanf} and
2056 @code{vsscanf} are also checked. Except in strictly conforming C
2057 standard modes, the X/Open function @code{strfmon} is also checked as
2058 are @code{printf_unlocked} and @code{fprintf_unlocked}.
2059 @xref{C Dialect Options,,Options Controlling C Dialect}.
2061 @item format_arg (@var{string-index})
2062 @cindex @code{format_arg} function attribute
2063 @opindex Wformat-nonliteral
2064 The @code{format_arg} attribute specifies that a function takes a format
2065 string for a @code{printf}, @code{scanf}, @code{strftime} or
2066 @code{strfmon} style function and modifies it (for example, to translate
2067 it into another language), so the result can be passed to a
2068 @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon} style
2069 function (with the remaining arguments to the format function the same
2070 as they would have been for the unmodified string). For example, the
2075 my_dgettext (char *my_domain, const char *my_format)
2076 __attribute__ ((format_arg (2)));
2080 causes the compiler to check the arguments in calls to a @code{printf},
2081 @code{scanf}, @code{strftime} or @code{strfmon} type function, whose
2082 format string argument is a call to the @code{my_dgettext} function, for
2083 consistency with the format string argument @code{my_format}. If the
2084 @code{format_arg} attribute had not been specified, all the compiler
2085 could tell in such calls to format functions would be that the format
2086 string argument is not constant; this would generate a warning when
2087 @option{-Wformat-nonliteral} is used, but the calls could not be checked
2088 without the attribute.
2090 The parameter @var{string-index} specifies which argument is the format
2091 string argument (starting from 1).
2093 The @code{format-arg} attribute allows you to identify your own
2094 functions which modify format strings, so that GCC can check the
2095 calls to @code{printf}, @code{scanf}, @code{strftime} or @code{strfmon}
2096 type function whose operands are a call to one of your own function.
2097 The compiler always treats @code{gettext}, @code{dgettext}, and
2098 @code{dcgettext} in this manner except when strict ISO C support is
2099 requested by @option{-ansi} or an appropriate @option{-std} option, or
2100 @option{-ffreestanding} is used. @xref{C Dialect Options,,Options
2101 Controlling C Dialect}.
2103 @item no_instrument_function
2104 @cindex @code{no_instrument_function} function attribute
2105 @opindex finstrument-functions
2106 If @option{-finstrument-functions} is given, profiling function calls will
2107 be generated at entry and exit of most user-compiled functions.
2108 Functions with this attribute will not be so instrumented.
2110 @item section ("@var{section-name}")
2111 @cindex @code{section} function attribute
2112 Normally, the compiler places the code it generates in the @code{text} section.
2113 Sometimes, however, you need additional sections, or you need certain
2114 particular functions to appear in special sections. The @code{section}
2115 attribute specifies that a function lives in a particular section.
2116 For example, the declaration:
2119 extern void foobar (void) __attribute__ ((section ("bar")));
2123 puts the function @code{foobar} in the @code{bar} section.
2125 Some file formats do not support arbitrary sections so the @code{section}
2126 attribute is not available on all platforms.
2127 If you need to map the entire contents of a module to a particular
2128 section, consider using the facilities of the linker instead.
2132 @cindex @code{constructor} function attribute
2133 @cindex @code{destructor} function attribute
2134 The @code{constructor} attribute causes the function to be called
2135 automatically before execution enters @code{main ()}. Similarly, the
2136 @code{destructor} attribute causes the function to be called
2137 automatically after @code{main ()} has completed or @code{exit ()} has
2138 been called. Functions with these attributes are useful for
2139 initializing data that will be used implicitly during the execution of
2142 These attributes are not currently implemented for Objective-C@.
2144 @cindex @code{unused} attribute.
2146 This attribute, attached to a function, means that the function is meant
2147 to be possibly unused. GCC will not produce a warning for this
2148 function. GNU C++ does not currently support this attribute as
2149 definitions without parameters are valid in C++.
2151 @cindex @code{used} attribute.
2153 This attribute, attached to a function, means that code must be emitted
2154 for the function even if it appears that the function is not referenced.
2155 This is useful, for example, when the function is referenced only in
2158 @cindex @code{deprecated} attribute.
2160 The @code{deprecated} attribute results in a warning if the function
2161 is used anywhere in the source file. This is useful when identifying
2162 functions that are expected to be removed in a future version of a
2163 program. The warning also includes the location of the declaration
2164 of the deprecated function, to enable users to easily find further
2165 information about why the function is deprecated, or what they should
2166 do instead. Note that the warnings only occurs for uses:
2169 int old_fn () __attribute__ ((deprecated));
2171 int (*fn_ptr)() = old_fn;
2174 results in a warning on line 3 but not line 2.
2176 The @code{deprecated} attribute can also be used for variables and
2177 types (@pxref{Variable Attributes}, @pxref{Type Attributes}.)
2180 @cindex @code{weak} attribute
2181 The @code{weak} attribute causes the declaration to be emitted as a weak
2182 symbol rather than a global. This is primarily useful in defining
2183 library functions which can be overridden in user code, though it can
2184 also be used with non-function declarations. Weak symbols are supported
2185 for ELF targets, and also for a.out targets when using the GNU assembler
2189 @cindex @code{malloc} attribute
2190 The @code{malloc} attribute is used to tell the compiler that a function
2191 may be treated as if it were the malloc function. The compiler assumes
2192 that calls to malloc result in a pointers that cannot alias anything.
2193 This will often improve optimization.
2195 @item alias ("@var{target}")
2196 @cindex @code{alias} attribute
2197 The @code{alias} attribute causes the declaration to be emitted as an
2198 alias for another symbol, which must be specified. For instance,
2201 void __f () @{ /* @r{Do something.} */; @}
2202 void f () __attribute__ ((weak, alias ("__f")));
2205 declares @samp{f} to be a weak alias for @samp{__f}. In C++, the
2206 mangled name for the target must be used.
2208 Not all target machines support this attribute.
2210 @item visibility ("@var{visibility_type}")
2211 @cindex @code{visibility} attribute
2212 The @code{visibility} attribute on ELF targets causes the declaration
2213 to be emitted with hidden, protected or internal visibility.
2216 void __attribute__ ((visibility ("protected")))
2217 f () @{ /* @r{Do something.} */; @}
2218 int i __attribute__ ((visibility ("hidden")));
2221 See the ELF gABI for complete details, but the short story is
2225 Hidden visibility indicates that the symbol will not be placed into
2226 the dynamic symbol table, so no other @dfn{module} (executable or
2227 shared library) can reference it directly.
2230 Protected visibility indicates that the symbol will be placed in the
2231 dynamic symbol table, but that references within the defining module
2232 will bind to the local symbol. That is, the symbol cannot be overridden
2236 Internal visibility is like hidden visibility, but with additional
2237 processor specific semantics. Unless otherwise specified by the psABI,
2238 gcc defines internal visibility to mean that the function is @emph{never}
2239 called from another module. Note that hidden symbols, while then cannot
2240 be referenced directly by other modules, can be referenced indirectly via
2241 function pointers. By indicating that a symbol cannot be called from
2242 outside the module, gcc may for instance omit the load of a PIC register
2243 since it is known that the calling function loaded the correct value.
2246 Not all ELF targets support this attribute.
2248 @item regparm (@var{number})
2249 @cindex functions that are passed arguments in registers on the 386
2250 On the Intel 386, the @code{regparm} attribute causes the compiler to
2251 pass up to @var{number} integer arguments in registers EAX,
2252 EDX, and ECX instead of on the stack. Functions that take a
2253 variable number of arguments will continue to be passed all of their
2254 arguments on the stack.
2257 @cindex functions that pop the argument stack on the 386
2258 On the Intel 386, the @code{stdcall} attribute causes the compiler to
2259 assume that the called function will pop off the stack space used to
2260 pass arguments, unless it takes a variable number of arguments.
2262 The PowerPC compiler for Windows NT currently ignores the @code{stdcall}
2266 @cindex functions that do pop the argument stack on the 386
2268 On the Intel 386, the @code{cdecl} attribute causes the compiler to
2269 assume that the calling function will pop off the stack space used to
2270 pass arguments. This is
2271 useful to override the effects of the @option{-mrtd} switch.
2273 The PowerPC compiler for Windows NT currently ignores the @code{cdecl}
2277 @cindex functions called via pointer on the RS/6000 and PowerPC
2278 On the RS/6000 and PowerPC, the @code{longcall} attribute causes the
2279 compiler to always call the function via a pointer, so that functions
2280 which reside further than 64 megabytes (67,108,864 bytes) from the
2281 current location can be called.
2283 @item long_call/short_call
2284 @cindex indirect calls on ARM
2285 This attribute allows to specify how to call a particular function on
2286 ARM@. Both attributes override the @option{-mlong-calls} (@pxref{ARM Options})
2287 command line switch and @code{#pragma long_calls} settings. The
2288 @code{long_call} attribute causes the compiler to always call the
2289 function by first loading its address into a register and then using the
2290 contents of that register. The @code{short_call} attribute always places
2291 the offset to the function from the call site into the @samp{BL}
2292 instruction directly.
2295 @cindex functions which are imported from a dll on PowerPC Windows NT
2296 On the PowerPC running Windows NT, the @code{dllimport} attribute causes
2297 the compiler to call the function via a global pointer to the function
2298 pointer that is set up by the Windows NT dll library. The pointer name
2299 is formed by combining @code{__imp_} and the function name.
2302 @cindex functions which are exported from a dll on PowerPC Windows NT
2303 On the PowerPC running Windows NT, the @code{dllexport} attribute causes
2304 the compiler to provide a global pointer to the function pointer, so
2305 that it can be called with the @code{dllimport} attribute. The pointer
2306 name is formed by combining @code{__imp_} and the function name.
2308 @item exception (@var{except-func} [, @var{except-arg}])
2309 @cindex functions which specify exception handling on PowerPC Windows NT
2310 On the PowerPC running Windows NT, the @code{exception} attribute causes
2311 the compiler to modify the structured exception table entry it emits for
2312 the declared function. The string or identifier @var{except-func} is
2313 placed in the third entry of the structured exception table. It
2314 represents a function, which is called by the exception handling
2315 mechanism if an exception occurs. If it was specified, the string or
2316 identifier @var{except-arg} is placed in the fourth entry of the
2317 structured exception table.
2319 @item function_vector
2320 @cindex calling functions through the function vector on the H8/300 processors
2321 Use this attribute on the H8/300 and H8/300H to indicate that the specified
2322 function should be called through the function vector. Calling a
2323 function through the function vector will reduce code size, however;
2324 the function vector has a limited size (maximum 128 entries on the H8/300
2325 and 64 entries on the H8/300H) and shares space with the interrupt vector.
2327 You must use GAS and GLD from GNU binutils version 2.7 or later for
2328 this attribute to work correctly.
2331 @cindex interrupt handler functions
2332 Use this attribute on the ARM, AVR, M32R/D and Xstormy16 ports to indicate
2333 that the specified function is an interrupt handler. The compiler will
2334 generate function entry and exit sequences suitable for use in an
2335 interrupt handler when this attribute is present.
2337 Note, interrupt handlers for the H8/300, H8/300H and SH processors can
2338 be specified via the @code{interrupt_handler} attribute.
2340 Note, on the AVR interrupts will be enabled inside the function.
2342 Note, for the ARM you can specify the kind of interrupt to be handled by
2343 adding an optional parameter to the interrupt attribute like this:
2346 void f () __attribute__ ((interrupt ("IRQ")));
2349 Permissible values for this parameter are: IRQ, FIQ, SWI, ABORT and UNDEF@.
2351 @item interrupt_handler
2352 @cindex interrupt handler functions on the H8/300 and SH processors
2353 Use this attribute on the H8/300, H8/300H and SH to indicate that the
2354 specified function is an interrupt handler. The compiler will generate
2355 function entry and exit sequences suitable for use in an interrupt
2356 handler when this attribute is present.
2359 Use this attribute on the SH to indicate an @code{interrupt_handler}
2360 function should switch to an alternate stack. It expects a string
2361 argument that names a global variable holding the address of the
2366 void f () __attribute__ ((interrupt_handler,
2367 sp_switch ("alt_stack")));
2371 Use this attribute on the SH for an @code{interrupt_handle} to return using
2372 @code{trapa} instead of @code{rte}. This attribute expects an integer
2373 argument specifying the trap number to be used.
2376 @cindex eight bit data on the H8/300 and H8/300H
2377 Use this attribute on the H8/300 and H8/300H to indicate that the specified
2378 variable should be placed into the eight bit data section.
2379 The compiler will generate more efficient code for certain operations
2380 on data in the eight bit data area. Note the eight bit data area is limited to
2383 You must use GAS and GLD from GNU binutils version 2.7 or later for
2384 this attribute to work correctly.
2387 @cindex tiny data section on the H8/300H
2388 Use this attribute on the H8/300H to indicate that the specified
2389 variable should be placed into the tiny data section.
2390 The compiler will generate more efficient code for loads and stores
2391 on data in the tiny data section. Note the tiny data area is limited to
2392 slightly under 32kbytes of data.
2395 @cindex signal handler functions on the AVR processors
2396 Use this attribute on the AVR to indicate that the specified
2397 function is an signal handler. The compiler will generate function
2398 entry and exit sequences suitable for use in an signal handler when this
2399 attribute is present. Interrupts will be disabled inside function.
2402 @cindex function without a prologue/epilogue code
2403 Use this attribute on the ARM or AVR ports to indicate that the specified
2404 function do not need prologue/epilogue sequences generated by the
2405 compiler. It is up to the programmer to provide these sequences.
2407 @item model (@var{model-name})
2408 @cindex function addressability on the M32R/D
2409 Use this attribute on the M32R/D to set the addressability of an object,
2410 and the code generated for a function.
2411 The identifier @var{model-name} is one of @code{small}, @code{medium},
2412 or @code{large}, representing each of the code models.
2414 Small model objects live in the lower 16MB of memory (so that their
2415 addresses can be loaded with the @code{ld24} instruction), and are
2416 callable with the @code{bl} instruction.
2418 Medium model objects may live anywhere in the 32-bit address space (the
2419 compiler will generate @code{seth/add3} instructions to load their addresses),
2420 and are callable with the @code{bl} instruction.
2422 Large model objects may live anywhere in the 32-bit address space (the
2423 compiler will generate @code{seth/add3} instructions to load their addresses),
2424 and may not be reachable with the @code{bl} instruction (the compiler will
2425 generate the much slower @code{seth/add3/jl} instruction sequence).
2429 You can specify multiple attributes in a declaration by separating them
2430 by commas within the double parentheses or by immediately following an
2431 attribute declaration with another attribute declaration.
2433 @cindex @code{#pragma}, reason for not using
2434 @cindex pragma, reason for not using
2435 Some people object to the @code{__attribute__} feature, suggesting that
2436 ISO C's @code{#pragma} should be used instead. At the time
2437 @code{__attribute__} was designed, there were two reasons for not doing
2442 It is impossible to generate @code{#pragma} commands from a macro.
2445 There is no telling what the same @code{#pragma} might mean in another
2449 These two reasons applied to almost any application that might have been
2450 proposed for @code{#pragma}. It was basically a mistake to use
2451 @code{#pragma} for @emph{anything}.
2453 The ISO C99 standard includes @code{_Pragma}, which now allows pragmas
2454 to be generated from macros. In addition, a @code{#pragma GCC}
2455 namespace is now in use for GCC-specific pragmas. However, it has been
2456 found convenient to use @code{__attribute__} to achieve a natural
2457 attachment of attributes to their corresponding declarations, whereas
2458 @code{#pragma GCC} is of use for constructs that do not naturally form
2459 part of the grammar. @xref{Other Directives,,Miscellaneous
2460 Preprocessing Directives, cpp, The C Preprocessor}.
2462 @node Attribute Syntax
2463 @section Attribute Syntax
2464 @cindex attribute syntax
2466 This section describes the syntax with which @code{__attribute__} may be
2467 used, and the constructs to which attribute specifiers bind, for the C
2468 language. Some details may vary for C++ and Objective-C@. Because of
2469 infelicities in the grammar for attributes, some forms described here
2470 may not be successfully parsed in all cases.
2472 There are some problems with the semantics of attributes in C++. For
2473 example, there are no manglings for attributes, although they may affect
2474 code generation, so problems may arise when attributed types are used in
2475 conjunction with templates or overloading. Similarly, @code{typeid}
2476 does not distinguish between types with different attributes. Support
2477 for attributes in C++ may be restricted in future to attributes on
2478 declarations only, but not on nested declarators.
2480 @xref{Function Attributes}, for details of the semantics of attributes
2481 applying to functions. @xref{Variable Attributes}, for details of the
2482 semantics of attributes applying to variables. @xref{Type Attributes},
2483 for details of the semantics of attributes applying to structure, union
2484 and enumerated types.
2486 An @dfn{attribute specifier} is of the form
2487 @code{__attribute__ ((@var{attribute-list}))}. An @dfn{attribute list}
2488 is a possibly empty comma-separated sequence of @dfn{attributes}, where
2489 each attribute is one of the following:
2493 Empty. Empty attributes are ignored.
2496 A word (which may be an identifier such as @code{unused}, or a reserved
2497 word such as @code{const}).
2500 A word, followed by, in parentheses, parameters for the attribute.
2501 These parameters take one of the following forms:
2505 An identifier. For example, @code{mode} attributes use this form.
2508 An identifier followed by a comma and a non-empty comma-separated list
2509 of expressions. For example, @code{format} attributes use this form.
2512 A possibly empty comma-separated list of expressions. For example,
2513 @code{format_arg} attributes use this form with the list being a single
2514 integer constant expression, and @code{alias} attributes use this form
2515 with the list being a single string constant.
2519 An @dfn{attribute specifier list} is a sequence of one or more attribute
2520 specifiers, not separated by any other tokens.
2522 An attribute specifier list may appear after the colon following a
2523 label, other than a @code{case} or @code{default} label. The only
2524 attribute it makes sense to use after a label is @code{unused}. This
2525 feature is intended for code generated by programs which contains labels
2526 that may be unused but which is compiled with @option{-Wall}. It would
2527 not normally be appropriate to use in it human-written code, though it
2528 could be useful in cases where the code that jumps to the label is
2529 contained within an @code{#ifdef} conditional.
2531 An attribute specifier list may appear as part of a @code{struct},
2532 @code{union} or @code{enum} specifier. It may go either immediately
2533 after the @code{struct}, @code{union} or @code{enum} keyword, or after
2534 the closing brace. It is ignored if the content of the structure, union
2535 or enumerated type is not defined in the specifier in which the
2536 attribute specifier list is used---that is, in usages such as
2537 @code{struct __attribute__((foo)) bar} with no following opening brace.
2538 Where attribute specifiers follow the closing brace, they are considered
2539 to relate to the structure, union or enumerated type defined, not to any
2540 enclosing declaration the type specifier appears in, and the type
2541 defined is not complete until after the attribute specifiers.
2542 @c Otherwise, there would be the following problems: a shift/reduce
2543 @c conflict between attributes binding the struct/union/enum and
2544 @c binding to the list of specifiers/qualifiers; and "aligned"
2545 @c attributes could use sizeof for the structure, but the size could be
2546 @c changed later by "packed" attributes.
2548 Otherwise, an attribute specifier appears as part of a declaration,
2549 counting declarations of unnamed parameters and type names, and relates
2550 to that declaration (which may be nested in another declaration, for
2551 example in the case of a parameter declaration), or to a particular declarator
2552 within a declaration. Where an
2553 attribute specifier is applied to a parameter declared as a function or
2554 an array, it should apply to the function or array rather than the
2555 pointer to which the parameter is implicitly converted, but this is not
2556 yet correctly implemented.
2558 Any list of specifiers and qualifiers at the start of a declaration may
2559 contain attribute specifiers, whether or not such a list may in that
2560 context contain storage class specifiers. (Some attributes, however,
2561 are essentially in the nature of storage class specifiers, and only make
2562 sense where storage class specifiers may be used; for example,
2563 @code{section}.) There is one necessary limitation to this syntax: the
2564 first old-style parameter declaration in a function definition cannot
2565 begin with an attribute specifier, because such an attribute applies to
2566 the function instead by syntax described below (which, however, is not
2567 yet implemented in this case). In some other cases, attribute
2568 specifiers are permitted by this grammar but not yet supported by the
2569 compiler. All attribute specifiers in this place relate to the
2570 declaration as a whole. In the obsolescent usage where a type of
2571 @code{int} is implied by the absence of type specifiers, such a list of
2572 specifiers and qualifiers may be an attribute specifier list with no
2573 other specifiers or qualifiers.
2575 An attribute specifier list may appear immediately before a declarator
2576 (other than the first) in a comma-separated list of declarators in a
2577 declaration of more than one identifier using a single list of
2578 specifiers and qualifiers. Such attribute specifiers apply
2579 only to the identifier before whose declarator they appear. For
2583 __attribute__((noreturn)) void d0 (void),
2584 __attribute__((format(printf, 1, 2))) d1 (const char *, ...),
2589 the @code{noreturn} attribute applies to all the functions
2590 declared; the @code{format} attribute only applies to @code{d1}.
2592 An attribute specifier list may appear immediately before the comma,
2593 @code{=} or semicolon terminating the declaration of an identifier other
2594 than a function definition. At present, such attribute specifiers apply
2595 to the declared object or function, but in future they may attach to the
2596 outermost adjacent declarator. In simple cases there is no difference,
2597 but, for example, in
2600 void (****f)(void) __attribute__((noreturn));
2604 at present the @code{noreturn} attribute applies to @code{f}, which
2605 causes a warning since @code{f} is not a function, but in future it may
2606 apply to the function @code{****f}. The precise semantics of what
2607 attributes in such cases will apply to are not yet specified. Where an
2608 assembler name for an object or function is specified (@pxref{Asm
2609 Labels}), at present the attribute must follow the @code{asm}
2610 specification; in future, attributes before the @code{asm} specification
2611 may apply to the adjacent declarator, and those after it to the declared
2614 An attribute specifier list may, in future, be permitted to appear after
2615 the declarator in a function definition (before any old-style parameter
2616 declarations or the function body).
2618 Attribute specifiers may be mixed with type qualifiers appearing inside
2619 the @code{[]} of a parameter array declarator, in the C99 construct by
2620 which such qualifiers are applied to the pointer to which the array is
2621 implicitly converted. Such attribute specifiers apply to the pointer,
2622 not to the array, but at present this is not implemented and they are
2625 An attribute specifier list may appear at the start of a nested
2626 declarator. At present, there are some limitations in this usage: the
2627 attributes correctly apply to the declarator, but for most individual
2628 attributes the semantics this implies are not implemented.
2629 When attribute specifiers follow the @code{*} of a pointer
2630 declarator, they may be mixed with any type qualifiers present.
2631 The following describes the formal semantics of this syntax. It will make the
2632 most sense if you are familiar with the formal specification of
2633 declarators in the ISO C standard.
2635 Consider (as in C99 subclause 6.7.5 paragraph 4) a declaration @code{T
2636 D1}, where @code{T} contains declaration specifiers that specify a type
2637 @var{Type} (such as @code{int}) and @code{D1} is a declarator that
2638 contains an identifier @var{ident}. The type specified for @var{ident}
2639 for derived declarators whose type does not include an attribute
2640 specifier is as in the ISO C standard.
2642 If @code{D1} has the form @code{( @var{attribute-specifier-list} D )},
2643 and the declaration @code{T D} specifies the type
2644 ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then
2645 @code{T D1} specifies the type ``@var{derived-declarator-type-list}
2646 @var{attribute-specifier-list} @var{Type}'' for @var{ident}.
2648 If @code{D1} has the form @code{*
2649 @var{type-qualifier-and-attribute-specifier-list} D}, and the
2650 declaration @code{T D} specifies the type
2651 ``@var{derived-declarator-type-list} @var{Type}'' for @var{ident}, then
2652 @code{T D1} specifies the type ``@var{derived-declarator-type-list}
2653 @var{type-qualifier-and-attribute-specifier-list} @var{Type}'' for
2659 void (__attribute__((noreturn)) ****f) (void);
2663 specifies the type ``pointer to pointer to pointer to pointer to
2664 non-returning function returning @code{void}''. As another example,
2667 char *__attribute__((aligned(8))) *f;
2671 specifies the type ``pointer to 8-byte-aligned pointer to @code{char}''.
2672 Note again that this does not work with most attributes; for example,
2673 the usage of @samp{aligned} and @samp{noreturn} attributes given above
2674 is not yet supported.
2676 For compatibility with existing code written for compiler versions that
2677 did not implement attributes on nested declarators, some laxity is
2678 allowed in the placing of attributes. If an attribute that only applies
2679 to types is applied to a declaration, it will be treated as applying to
2680 the type of that declaration. If an attribute that only applies to
2681 declarations is applied to the type of a declaration, it will be treated
2682 as applying to that declaration; and, for compatibility with code
2683 placing the attributes immediately before the identifier declared, such
2684 an attribute applied to a function return type will be treated as
2685 applying to the function type, and such an attribute applied to an array
2686 element type will be treated as applying to the array type. If an
2687 attribute that only applies to function types is applied to a
2688 pointer-to-function type, it will be treated as applying to the pointer
2689 target type; if such an attribute is applied to a function return type
2690 that is not a pointer-to-function type, it will be treated as applying
2691 to the function type.
2693 @node Function Prototypes
2694 @section Prototypes and Old-Style Function Definitions
2695 @cindex function prototype declarations
2696 @cindex old-style function definitions
2697 @cindex promotion of formal parameters
2699 GNU C extends ISO C to allow a function prototype to override a later
2700 old-style non-prototype definition. Consider the following example:
2703 /* @r{Use prototypes unless the compiler is old-fashioned.} */
2710 /* @r{Prototype function declaration.} */
2711 int isroot P((uid_t));
2713 /* @r{Old-style function definition.} */
2715 isroot (x) /* ??? lossage here ??? */
2722 Suppose the type @code{uid_t} happens to be @code{short}. ISO C does
2723 not allow this example, because subword arguments in old-style
2724 non-prototype definitions are promoted. Therefore in this example the
2725 function definition's argument is really an @code{int}, which does not
2726 match the prototype argument type of @code{short}.
2728 This restriction of ISO C makes it hard to write code that is portable
2729 to traditional C compilers, because the programmer does not know
2730 whether the @code{uid_t} type is @code{short}, @code{int}, or
2731 @code{long}. Therefore, in cases like these GNU C allows a prototype
2732 to override a later old-style definition. More precisely, in GNU C, a
2733 function prototype argument type overrides the argument type specified
2734 by a later old-style definition if the former type is the same as the
2735 latter type before promotion. Thus in GNU C the above example is
2736 equivalent to the following:
2749 GNU C++ does not support old-style function definitions, so this
2750 extension is irrelevant.
2753 @section C++ Style Comments
2755 @cindex C++ comments
2756 @cindex comments, C++ style
2758 In GNU C, you may use C++ style comments, which start with @samp{//} and
2759 continue until the end of the line. Many other C implementations allow
2760 such comments, and they are included in the 1999 C standard. However,
2761 C++ style comments are not recognized if you specify an @option{-std}
2762 option specifying a version of ISO C before C99, or @option{-ansi}
2763 (equivalent to @option{-std=c89}).
2766 @section Dollar Signs in Identifier Names
2768 @cindex dollar signs in identifier names
2769 @cindex identifier names, dollar signs in
2771 In GNU C, you may normally use dollar signs in identifier names.
2772 This is because many traditional C implementations allow such identifiers.
2773 However, dollar signs in identifiers are not supported on a few target
2774 machines, typically because the target assembler does not allow them.
2776 @node Character Escapes
2777 @section The Character @key{ESC} in Constants
2779 You can use the sequence @samp{\e} in a string or character constant to
2780 stand for the ASCII character @key{ESC}.
2783 @section Inquiring on Alignment of Types or Variables
2785 @cindex type alignment
2786 @cindex variable alignment
2788 The keyword @code{__alignof__} allows you to inquire about how an object
2789 is aligned, or the minimum alignment usually required by a type. Its
2790 syntax is just like @code{sizeof}.
2792 For example, if the target machine requires a @code{double} value to be
2793 aligned on an 8-byte boundary, then @code{__alignof__ (double)} is 8.
2794 This is true on many RISC machines. On more traditional machine
2795 designs, @code{__alignof__ (double)} is 4 or even 2.
2797 Some machines never actually require alignment; they allow reference to any
2798 data type even at an odd addresses. For these machines, @code{__alignof__}
2799 reports the @emph{recommended} alignment of a type.
2801 If the operand of @code{__alignof__} is an lvalue rather than a type,
2802 its value is the required alignment for its type, taking into account
2803 any minimum alignment specified with GCC's @code{__attribute__}
2804 extension (@pxref{Variable Attributes}). For example, after this
2808 struct foo @{ int x; char y; @} foo1;
2812 the value of @code{__alignof__ (foo1.y)} is 1, even though its actual
2813 alignment is probably 2 or 4, the same as @code{__alignof__ (int)}.
2815 It is an error to ask for the alignment of an incomplete type.
2817 @node Variable Attributes
2818 @section Specifying Attributes of Variables
2819 @cindex attribute of variables
2820 @cindex variable attributes
2822 The keyword @code{__attribute__} allows you to specify special
2823 attributes of variables or structure fields. This keyword is followed
2824 by an attribute specification inside double parentheses. Ten
2825 attributes are currently defined for variables: @code{aligned},
2826 @code{mode}, @code{nocommon}, @code{packed}, @code{section},
2827 @code{transparent_union}, @code{unused}, @code{deprecated},
2828 @code{vector_size}, and @code{weak}. Some other attributes are defined
2829 for variables on particular target systems. Other attributes are
2830 available for functions (@pxref{Function Attributes}) and for types
2831 (@pxref{Type Attributes}). Other front ends might define more
2832 attributes (@pxref{C++ Extensions,,Extensions to the C++ Language}).
2834 You may also specify attributes with @samp{__} preceding and following
2835 each keyword. This allows you to use them in header files without
2836 being concerned about a possible macro of the same name. For example,
2837 you may use @code{__aligned__} instead of @code{aligned}.
2839 @xref{Attribute Syntax}, for details of the exact syntax for using
2843 @cindex @code{aligned} attribute
2844 @item aligned (@var{alignment})
2845 This attribute specifies a minimum alignment for the variable or
2846 structure field, measured in bytes. For example, the declaration:
2849 int x __attribute__ ((aligned (16))) = 0;
2853 causes the compiler to allocate the global variable @code{x} on a
2854 16-byte boundary. On a 68040, this could be used in conjunction with
2855 an @code{asm} expression to access the @code{move16} instruction which
2856 requires 16-byte aligned operands.
2858 You can also specify the alignment of structure fields. For example, to
2859 create a double-word aligned @code{int} pair, you could write:
2862 struct foo @{ int x[2] __attribute__ ((aligned (8))); @};
2866 This is an alternative to creating a union with a @code{double} member
2867 that forces the union to be double-word aligned.
2869 As in the preceding examples, you can explicitly specify the alignment
2870 (in bytes) that you wish the compiler to use for a given variable or
2871 structure field. Alternatively, you can leave out the alignment factor
2872 and just ask the compiler to align a variable or field to the maximum
2873 useful alignment for the target machine you are compiling for. For
2874 example, you could write:
2877 short array[3] __attribute__ ((aligned));
2880 Whenever you leave out the alignment factor in an @code{aligned} attribute
2881 specification, the compiler automatically sets the alignment for the declared
2882 variable or field to the largest alignment which is ever used for any data
2883 type on the target machine you are compiling for. Doing this can often make
2884 copy operations more efficient, because the compiler can use whatever
2885 instructions copy the biggest chunks of memory when performing copies to
2886 or from the variables or fields that you have aligned this way.
2888 The @code{aligned} attribute can only increase the alignment; but you
2889 can decrease it by specifying @code{packed} as well. See below.
2891 Note that the effectiveness of @code{aligned} attributes may be limited
2892 by inherent limitations in your linker. On many systems, the linker is
2893 only able to arrange for variables to be aligned up to a certain maximum
2894 alignment. (For some linkers, the maximum supported alignment may
2895 be very very small.) If your linker is only able to align variables
2896 up to a maximum of 8 byte alignment, then specifying @code{aligned(16)}
2897 in an @code{__attribute__} will still only provide you with 8 byte
2898 alignment. See your linker documentation for further information.
2900 @item mode (@var{mode})
2901 @cindex @code{mode} attribute
2902 This attribute specifies the data type for the declaration---whichever
2903 type corresponds to the mode @var{mode}. This in effect lets you
2904 request an integer or floating point type according to its width.
2906 You may also specify a mode of @samp{byte} or @samp{__byte__} to
2907 indicate the mode corresponding to a one-byte integer, @samp{word} or
2908 @samp{__word__} for the mode of a one-word integer, and @samp{pointer}
2909 or @samp{__pointer__} for the mode used to represent pointers.
2912 @cindex @code{nocommon} attribute
2914 This attribute specifies requests GCC not to place a variable
2915 ``common'' but instead to allocate space for it directly. If you
2916 specify the @option{-fno-common} flag, GCC will do this for all
2919 Specifying the @code{nocommon} attribute for a variable provides an
2920 initialization of zeros. A variable may only be initialized in one
2924 @cindex @code{packed} attribute
2925 The @code{packed} attribute specifies that a variable or structure field
2926 should have the smallest possible alignment---one byte for a variable,
2927 and one bit for a field, unless you specify a larger value with the
2928 @code{aligned} attribute.
2930 Here is a structure in which the field @code{x} is packed, so that it
2931 immediately follows @code{a}:
2937 int x[2] __attribute__ ((packed));
2941 @item section ("@var{section-name}")
2942 @cindex @code{section} variable attribute
2943 Normally, the compiler places the objects it generates in sections like
2944 @code{data} and @code{bss}. Sometimes, however, you need additional sections,
2945 or you need certain particular variables to appear in special sections,
2946 for example to map to special hardware. The @code{section}
2947 attribute specifies that a variable (or function) lives in a particular
2948 section. For example, this small program uses several specific section names:
2951 struct duart a __attribute__ ((section ("DUART_A"))) = @{ 0 @};
2952 struct duart b __attribute__ ((section ("DUART_B"))) = @{ 0 @};
2953 char stack[10000] __attribute__ ((section ("STACK"))) = @{ 0 @};
2954 int init_data __attribute__ ((section ("INITDATA"))) = 0;
2958 /* Initialize stack pointer */
2959 init_sp (stack + sizeof (stack));
2961 /* Initialize initialized data */
2962 memcpy (&init_data, &data, &edata - &data);
2964 /* Turn on the serial ports */
2971 Use the @code{section} attribute with an @emph{initialized} definition
2972 of a @emph{global} variable, as shown in the example. GCC issues
2973 a warning and otherwise ignores the @code{section} attribute in
2974 uninitialized variable declarations.
2976 You may only use the @code{section} attribute with a fully initialized
2977 global definition because of the way linkers work. The linker requires
2978 each object be defined once, with the exception that uninitialized
2979 variables tentatively go in the @code{common} (or @code{bss}) section
2980 and can be multiply ``defined''. You can force a variable to be
2981 initialized with the @option{-fno-common} flag or the @code{nocommon}
2984 Some file formats do not support arbitrary sections so the @code{section}
2985 attribute is not available on all platforms.
2986 If you need to map the entire contents of a module to a particular
2987 section, consider using the facilities of the linker instead.
2990 @cindex @code{shared} variable attribute
2991 On Windows NT, in addition to putting variable definitions in a named
2992 section, the section can also be shared among all running copies of an
2993 executable or DLL@. For example, this small program defines shared data
2994 by putting it in a named section @code{shared} and marking the section
2998 int foo __attribute__((section ("shared"), shared)) = 0;
3003 /* Read and write foo. All running
3004 copies see the same value. */
3010 You may only use the @code{shared} attribute along with @code{section}
3011 attribute with a fully initialized global definition because of the way
3012 linkers work. See @code{section} attribute for more information.
3014 The @code{shared} attribute is only available on Windows NT@.
3016 @item transparent_union
3017 This attribute, attached to a function parameter which is a union, means
3018 that the corresponding argument may have the type of any union member,
3019 but the argument is passed as if its type were that of the first union
3020 member. For more details see @xref{Type Attributes}. You can also use
3021 this attribute on a @code{typedef} for a union data type; then it
3022 applies to all function parameters with that type.
3025 This attribute, attached to a variable, means that the variable is meant
3026 to be possibly unused. GCC will not produce a warning for this
3030 The @code{deprecated} attribute results in a warning if the variable
3031 is used anywhere in the source file. This is useful when identifying
3032 variables that are expected to be removed in a future version of a
3033 program. The warning also includes the location of the declaration
3034 of the deprecated variable, to enable users to easily find further
3035 information about why the variable is deprecated, or what they should
3036 do instead. Note that the warnings only occurs for uses:
3039 extern int old_var __attribute__ ((deprecated));
3041 int new_fn () @{ return old_var; @}
3044 results in a warning on line 3 but not line 2.
3046 The @code{deprecated} attribute can also be used for functions and
3047 types (@pxref{Function Attributes}, @pxref{Type Attributes}.)
3049 @item vector_size (@var{bytes})
3050 This attribute specifies the vector size for the variable, measured in
3051 bytes. For example, the declaration:
3054 int foo __attribute__ ((vector_size (16)));
3058 causes the compiler to set the mode for @code{foo}, to be 16 bytes,
3059 divided into @code{int} sized units. Assuming a 32-bit int (a vector of
3060 4 units of 4 bytes), the corresponding mode of @code{foo} will be V4SI@.
3062 This attribute is only applicable to integral and float scalars,
3063 although arrays, pointers, and function return values are allowed in
3064 conjunction with this construct.
3066 Aggregates with this attribute are invalid, even if they are of the same
3067 size as a corresponding scalar. For example, the declaration:
3070 struct S @{ int a; @};
3071 struct S __attribute__ ((vector_size (16))) foo;
3075 is invalid even if the size of the structure is the same as the size of
3079 The @code{weak} attribute is described in @xref{Function Attributes}.
3081 @item model (@var{model-name})
3082 @cindex variable addressability on the M32R/D
3083 Use this attribute on the M32R/D to set the addressability of an object.
3084 The identifier @var{model-name} is one of @code{small}, @code{medium},
3085 or @code{large}, representing each of the code models.
3087 Small model objects live in the lower 16MB of memory (so that their
3088 addresses can be loaded with the @code{ld24} instruction).
3090 Medium and large model objects may live anywhere in the 32-bit address space
3091 (the compiler will generate @code{seth/add3} instructions to load their
3096 To specify multiple attributes, separate them by commas within the
3097 double parentheses: for example, @samp{__attribute__ ((aligned (16),
3100 @node Type Attributes
3101 @section Specifying Attributes of Types
3102 @cindex attribute of types
3103 @cindex type attributes
3105 The keyword @code{__attribute__} allows you to specify special
3106 attributes of @code{struct} and @code{union} types when you define such
3107 types. This keyword is followed by an attribute specification inside
3108 double parentheses. Five attributes are currently defined for types:
3109 @code{aligned}, @code{packed}, @code{transparent_union}, @code{unused},
3110 and @code{deprecated}. Other attributes are defined for functions
3111 (@pxref{Function Attributes}) and for variables (@pxref{Variable Attributes}).
3113 You may also specify any one of these attributes with @samp{__}
3114 preceding and following its keyword. This allows you to use these
3115 attributes in header files without being concerned about a possible
3116 macro of the same name. For example, you may use @code{__aligned__}
3117 instead of @code{aligned}.
3119 You may specify the @code{aligned} and @code{transparent_union}
3120 attributes either in a @code{typedef} declaration or just past the
3121 closing curly brace of a complete enum, struct or union type
3122 @emph{definition} and the @code{packed} attribute only past the closing
3123 brace of a definition.
3125 You may also specify attributes between the enum, struct or union
3126 tag and the name of the type rather than after the closing brace.
3128 @xref{Attribute Syntax}, for details of the exact syntax for using
3132 @cindex @code{aligned} attribute
3133 @item aligned (@var{alignment})
3134 This attribute specifies a minimum alignment (in bytes) for variables
3135 of the specified type. For example, the declarations:
3138 struct S @{ short f[3]; @} __attribute__ ((aligned (8)));
3139 typedef int more_aligned_int __attribute__ ((aligned (8)));
3143 force the compiler to insure (as far as it can) that each variable whose
3144 type is @code{struct S} or @code{more_aligned_int} will be allocated and
3145 aligned @emph{at least} on a 8-byte boundary. On a Sparc, having all
3146 variables of type @code{struct S} aligned to 8-byte boundaries allows
3147 the compiler to use the @code{ldd} and @code{std} (doubleword load and
3148 store) instructions when copying one variable of type @code{struct S} to
3149 another, thus improving run-time efficiency.
3151 Note that the alignment of any given @code{struct} or @code{union} type
3152 is required by the ISO C standard to be at least a perfect multiple of
3153 the lowest common multiple of the alignments of all of the members of
3154 the @code{struct} or @code{union} in question. This means that you @emph{can}
3155 effectively adjust the alignment of a @code{struct} or @code{union}
3156 type by attaching an @code{aligned} attribute to any one of the members
3157 of such a type, but the notation illustrated in the example above is a
3158 more obvious, intuitive, and readable way to request the compiler to
3159 adjust the alignment of an entire @code{struct} or @code{union} type.
3161 As in the preceding example, you can explicitly specify the alignment
3162 (in bytes) that you wish the compiler to use for a given @code{struct}
3163 or @code{union} type. Alternatively, you can leave out the alignment factor
3164 and just ask the compiler to align a type to the maximum
3165 useful alignment for the target machine you are compiling for. For
3166 example, you could write:
3169 struct S @{ short f[3]; @} __attribute__ ((aligned));
3172 Whenever you leave out the alignment factor in an @code{aligned}
3173 attribute specification, the compiler automatically sets the alignment
3174 for the type to the largest alignment which is ever used for any data
3175 type on the target machine you are compiling for. Doing this can often
3176 make copy operations more efficient, because the compiler can use
3177 whatever instructions copy the biggest chunks of memory when performing
3178 copies to or from the variables which have types that you have aligned
3181 In the example above, if the size of each @code{short} is 2 bytes, then
3182 the size of the entire @code{struct S} type is 6 bytes. The smallest
3183 power of two which is greater than or equal to that is 8, so the
3184 compiler sets the alignment for the entire @code{struct S} type to 8
3187 Note that although you can ask the compiler to select a time-efficient
3188 alignment for a given type and then declare only individual stand-alone
3189 objects of that type, the compiler's ability to select a time-efficient
3190 alignment is primarily useful only when you plan to create arrays of
3191 variables having the relevant (efficiently aligned) type. If you
3192 declare or use arrays of variables of an efficiently-aligned type, then
3193 it is likely that your program will also be doing pointer arithmetic (or
3194 subscripting, which amounts to the same thing) on pointers to the
3195 relevant type, and the code that the compiler generates for these
3196 pointer arithmetic operations will often be more efficient for
3197 efficiently-aligned types than for other types.
3199 The @code{aligned} attribute can only increase the alignment; but you
3200 can decrease it by specifying @code{packed} as well. See below.
3202 Note that the effectiveness of @code{aligned} attributes may be limited
3203 by inherent limitations in your linker. On many systems, the linker is
3204 only able to arrange for variables to be aligned up to a certain maximum
3205 alignment. (For some linkers, the maximum supported alignment may
3206 be very very small.) If your linker is only able to align variables
3207 up to a maximum of 8 byte alignment, then specifying @code{aligned(16)}
3208 in an @code{__attribute__} will still only provide you with 8 byte
3209 alignment. See your linker documentation for further information.
3212 This attribute, attached to an @code{enum}, @code{struct}, or
3213 @code{union} type definition, specified that the minimum required memory
3214 be used to represent the type.
3216 @opindex fshort-enums
3217 Specifying this attribute for @code{struct} and @code{union} types is
3218 equivalent to specifying the @code{packed} attribute on each of the
3219 structure or union members. Specifying the @option{-fshort-enums}
3220 flag on the line is equivalent to specifying the @code{packed}
3221 attribute on all @code{enum} definitions.
3223 You may only specify this attribute after a closing curly brace on an
3224 @code{enum} definition, not in a @code{typedef} declaration, unless that
3225 declaration also contains the definition of the @code{enum}.
3227 @item transparent_union
3228 This attribute, attached to a @code{union} type definition, indicates
3229 that any function parameter having that union type causes calls to that
3230 function to be treated in a special way.
3232 First, the argument corresponding to a transparent union type can be of
3233 any type in the union; no cast is required. Also, if the union contains
3234 a pointer type, the corresponding argument can be a null pointer
3235 constant or a void pointer expression; and if the union contains a void
3236 pointer type, the corresponding argument can be any pointer expression.
3237 If the union member type is a pointer, qualifiers like @code{const} on
3238 the referenced type must be respected, just as with normal pointer
3241 Second, the argument is passed to the function using the calling
3242 conventions of first member of the transparent union, not the calling
3243 conventions of the union itself. All members of the union must have the
3244 same machine representation; this is necessary for this argument passing
3247 Transparent unions are designed for library functions that have multiple
3248 interfaces for compatibility reasons. For example, suppose the
3249 @code{wait} function must accept either a value of type @code{int *} to
3250 comply with Posix, or a value of type @code{union wait *} to comply with
3251 the 4.1BSD interface. If @code{wait}'s parameter were @code{void *},
3252 @code{wait} would accept both kinds of arguments, but it would also
3253 accept any other pointer type and this would make argument type checking
3254 less useful. Instead, @code{<sys/wait.h>} might define the interface
3262 @} wait_status_ptr_t __attribute__ ((__transparent_union__));
3264 pid_t wait (wait_status_ptr_t);
3267 This interface allows either @code{int *} or @code{union wait *}
3268 arguments to be passed, using the @code{int *} calling convention.
3269 The program can call @code{wait} with arguments of either type:
3272 int w1 () @{ int w; return wait (&w); @}
3273 int w2 () @{ union wait w; return wait (&w); @}
3276 With this interface, @code{wait}'s implementation might look like this:
3279 pid_t wait (wait_status_ptr_t p)
3281 return waitpid (-1, p.__ip, 0);
3286 When attached to a type (including a @code{union} or a @code{struct}),
3287 this attribute means that variables of that type are meant to appear
3288 possibly unused. GCC will not produce a warning for any variables of
3289 that type, even if the variable appears to do nothing. This is often
3290 the case with lock or thread classes, which are usually defined and then
3291 not referenced, but contain constructors and destructors that have
3292 nontrivial bookkeeping functions.
3295 The @code{deprecated} attribute results in a warning if the type
3296 is used anywhere in the source file. This is useful when identifying
3297 types that are expected to be removed in a future version of a program.
3298 If possible, the warning also includes the location of the declaration
3299 of the deprecated type, to enable users to easily find further
3300 information about why the type is deprecated, or what they should do
3301 instead. Note that the warnings only occur for uses and then only
3302 if the type is being applied to an identifier that itself is not being
3303 declared as deprecated.
3306 typedef int T1 __attribute__ ((deprecated));
3310 typedef T1 T3 __attribute__ ((deprecated));
3311 T3 z __attribute__ ((deprecated));
3314 results in a warning on line 2 and 3 but not lines 4, 5, or 6. No
3315 warning is issued for line 4 because T2 is not explicitly
3316 deprecated. Line 5 has no warning because T3 is explicitly
3317 deprecated. Similarly for line 6.
3319 The @code{deprecated} attribute can also be used for functions and
3320 variables (@pxref{Function Attributes}, @pxref{Variable Attributes}.)
3324 To specify multiple attributes, separate them by commas within the
3325 double parentheses: for example, @samp{__attribute__ ((aligned (16),
3329 @section An Inline Function is As Fast As a Macro
3330 @cindex inline functions
3331 @cindex integrating function code
3333 @cindex macros, inline alternative
3335 By declaring a function @code{inline}, you can direct GCC to
3336 integrate that function's code into the code for its callers. This
3337 makes execution faster by eliminating the function-call overhead; in
3338 addition, if any of the actual argument values are constant, their known
3339 values may permit simplifications at compile time so that not all of the
3340 inline function's code needs to be included. The effect on code size is
3341 less predictable; object code may be larger or smaller with function
3342 inlining, depending on the particular case. Inlining of functions is an
3343 optimization and it really ``works'' only in optimizing compilation. If
3344 you don't use @option{-O}, no function is really inline.
3346 Inline functions are included in the ISO C99 standard, but there are
3347 currently substantial differences between what GCC implements and what
3348 the ISO C99 standard requires.
3350 To declare a function inline, use the @code{inline} keyword in its
3351 declaration, like this:
3361 (If you are writing a header file to be included in ISO C programs, write
3362 @code{__inline__} instead of @code{inline}. @xref{Alternate Keywords}.)
3363 You can also make all ``simple enough'' functions inline with the option
3364 @option{-finline-functions}.
3367 Note that certain usages in a function definition can make it unsuitable
3368 for inline substitution. Among these usages are: use of varargs, use of
3369 alloca, use of variable sized data types (@pxref{Variable Length}),
3370 use of computed goto (@pxref{Labels as Values}), use of nonlocal goto,
3371 and nested functions (@pxref{Nested Functions}). Using @option{-Winline}
3372 will warn when a function marked @code{inline} could not be substituted,
3373 and will give the reason for the failure.
3375 Note that in C and Objective-C, unlike C++, the @code{inline} keyword
3376 does not affect the linkage of the function.
3378 @cindex automatic @code{inline} for C++ member fns
3379 @cindex @code{inline} automatic for C++ member fns
3380 @cindex member fns, automatically @code{inline}
3381 @cindex C++ member fns, automatically @code{inline}
3382 @opindex fno-default-inline
3383 GCC automatically inlines member functions defined within the class
3384 body of C++ programs even if they are not explicitly declared
3385 @code{inline}. (You can override this with @option{-fno-default-inline};
3386 @pxref{C++ Dialect Options,,Options Controlling C++ Dialect}.)
3388 @cindex inline functions, omission of
3389 @opindex fkeep-inline-functions
3390 When a function is both inline and @code{static}, if all calls to the
3391 function are integrated into the caller, and the function's address is
3392 never used, then the function's own assembler code is never referenced.
3393 In this case, GCC does not actually output assembler code for the
3394 function, unless you specify the option @option{-fkeep-inline-functions}.
3395 Some calls cannot be integrated for various reasons (in particular,
3396 calls that precede the function's definition cannot be integrated, and
3397 neither can recursive calls within the definition). If there is a
3398 nonintegrated call, then the function is compiled to assembler code as
3399 usual. The function must also be compiled as usual if the program
3400 refers to its address, because that can't be inlined.
3402 @cindex non-static inline function
3403 When an inline function is not @code{static}, then the compiler must assume
3404 that there may be calls from other source files; since a global symbol can
3405 be defined only once in any program, the function must not be defined in
3406 the other source files, so the calls therein cannot be integrated.
3407 Therefore, a non-@code{static} inline function is always compiled on its
3408 own in the usual fashion.
3410 If you specify both @code{inline} and @code{extern} in the function
3411 definition, then the definition is used only for inlining. In no case
3412 is the function compiled on its own, not even if you refer to its
3413 address explicitly. Such an address becomes an external reference, as
3414 if you had only declared the function, and had not defined it.
3416 This combination of @code{inline} and @code{extern} has almost the
3417 effect of a macro. The way to use it is to put a function definition in
3418 a header file with these keywords, and put another copy of the
3419 definition (lacking @code{inline} and @code{extern}) in a library file.
3420 The definition in the header file will cause most calls to the function
3421 to be inlined. If any uses of the function remain, they will refer to
3422 the single copy in the library.
3424 For future compatibility with when GCC implements ISO C99 semantics for
3425 inline functions, it is best to use @code{static inline} only. (The
3426 existing semantics will remain available when @option{-std=gnu89} is
3427 specified, but eventually the default will be @option{-std=gnu99} and
3428 that will implement the C99 semantics, though it does not do so yet.)
3430 GCC does not inline any functions when not optimizing unless you specify
3431 the @samp{always_inline} attribute for the function, like this:
3435 inline void foo (const char) __attribute__((always_inline));
3439 @section Assembler Instructions with C Expression Operands
3440 @cindex extended @code{asm}
3441 @cindex @code{asm} expressions
3442 @cindex assembler instructions
3445 In an assembler instruction using @code{asm}, you can specify the
3446 operands of the instruction using C expressions. This means you need not
3447 guess which registers or memory locations will contain the data you want
3450 You must specify an assembler instruction template much like what
3451 appears in a machine description, plus an operand constraint string for
3454 For example, here is how to use the 68881's @code{fsinx} instruction:
3457 asm ("fsinx %1,%0" : "=f" (result) : "f" (angle));
3461 Here @code{angle} is the C expression for the input operand while
3462 @code{result} is that of the output operand. Each has @samp{"f"} as its
3463 operand constraint, saying that a floating point register is required.
3464 The @samp{=} in @samp{=f} indicates that the operand is an output; all
3465 output operands' constraints must use @samp{=}. The constraints use the
3466 same language used in the machine description (@pxref{Constraints}).
3468 Each operand is described by an operand-constraint string followed by
3469 the C expression in parentheses. A colon separates the assembler
3470 template from the first output operand and another separates the last
3471 output operand from the first input, if any. Commas separate the
3472 operands within each group. The total number of operands is currently
3473 limited to 30; this limitation may be lifted in some future version of
3476 If there are no output operands but there are input operands, you must
3477 place two consecutive colons surrounding the place where the output
3480 As of GCC version 3.1, it is also possible to specify input and output
3481 operands using symbolic names which can be referenced within the
3482 assembler code. These names are specified inside square brackets
3483 preceding the constraint string, and can be referenced inside the
3484 assembler code using @code{%[@var{name}]} instead of a percentage sign
3485 followed by the operand number. Using named operands the above example
3489 asm ("fsinx %[angle],%[output]"
3490 : [output] "=f" (result)
3491 : [angle] "f" (angle));
3495 Note that the symbolic operand names have no relation whatsoever to
3496 other C identifiers. You may use any name you like, even those of
3497 existing C symbols, but must ensure that no two operands within the same
3498 assembler construct use the same symbolic name.
3500 Output operand expressions must be lvalues; the compiler can check this.
3501 The input operands need not be lvalues. The compiler cannot check
3502 whether the operands have data types that are reasonable for the
3503 instruction being executed. It does not parse the assembler instruction
3504 template and does not know what it means or even whether it is valid
3505 assembler input. The extended @code{asm} feature is most often used for
3506 machine instructions the compiler itself does not know exist. If
3507 the output expression cannot be directly addressed (for example, it is a
3508 bit-field), your constraint must allow a register. In that case, GCC
3509 will use the register as the output of the @code{asm}, and then store
3510 that register into the output.
3512 The ordinary output operands must be write-only; GCC will assume that
3513 the values in these operands before the instruction are dead and need
3514 not be generated. Extended asm supports input-output or read-write
3515 operands. Use the constraint character @samp{+} to indicate such an
3516 operand and list it with the output operands.
3518 When the constraints for the read-write operand (or the operand in which
3519 only some of the bits are to be changed) allows a register, you may, as
3520 an alternative, logically split its function into two separate operands,
3521 one input operand and one write-only output operand. The connection
3522 between them is expressed by constraints which say they need to be in
3523 the same location when the instruction executes. You can use the same C
3524 expression for both operands, or different expressions. For example,
3525 here we write the (fictitious) @samp{combine} instruction with
3526 @code{bar} as its read-only source operand and @code{foo} as its
3527 read-write destination:
3530 asm ("combine %2,%0" : "=r" (foo) : "0" (foo), "g" (bar));
3534 The constraint @samp{"0"} for operand 1 says that it must occupy the
3535 same location as operand 0. A number in constraint is allowed only in
3536 an input operand and it must refer to an output operand.
3538 Only a number in the constraint can guarantee that one operand will be in
3539 the same place as another. The mere fact that @code{foo} is the value
3540 of both operands is not enough to guarantee that they will be in the
3541 same place in the generated assembler code. The following would not
3545 asm ("combine %2,%0" : "=r" (foo) : "r" (foo), "g" (bar));
3548 Various optimizations or reloading could cause operands 0 and 1 to be in
3549 different registers; GCC knows no reason not to do so. For example, the
3550 compiler might find a copy of the value of @code{foo} in one register and
3551 use it for operand 1, but generate the output operand 0 in a different
3552 register (copying it afterward to @code{foo}'s own address). Of course,
3553 since the register for operand 1 is not even mentioned in the assembler
3554 code, the result will not work, but GCC can't tell that.
3556 As of GCC version 3.1, one may write @code{[@var{name}]} instead of
3557 the operand number for a matching constraint. For example:
3560 asm ("cmoveq %1,%2,%[result]"
3561 : [result] "=r"(result)
3562 : "r" (test), "r"(new), "[result]"(old));
3565 Some instructions clobber specific hard registers. To describe this,
3566 write a third colon after the input operands, followed by the names of
3567 the clobbered hard registers (given as strings). Here is a realistic
3568 example for the VAX:
3571 asm volatile ("movc3 %0,%1,%2"
3573 : "g" (from), "g" (to), "g" (count)
3574 : "r0", "r1", "r2", "r3", "r4", "r5");
3577 You may not write a clobber description in a way that overlaps with an
3578 input or output operand. For example, you may not have an operand
3579 describing a register class with one member if you mention that register
3580 in the clobber list. There is no way for you to specify that an input
3581 operand is modified without also specifying it as an output
3582 operand. Note that if all the output operands you specify are for this
3583 purpose (and hence unused), you will then also need to specify
3584 @code{volatile} for the @code{asm} construct, as described below, to
3585 prevent GCC from deleting the @code{asm} statement as unused.
3587 If you refer to a particular hardware register from the assembler code,
3588 you will probably have to list the register after the third colon to
3589 tell the compiler the register's value is modified. In some assemblers,
3590 the register names begin with @samp{%}; to produce one @samp{%} in the
3591 assembler code, you must write @samp{%%} in the input.
3593 If your assembler instruction can alter the condition code register, add
3594 @samp{cc} to the list of clobbered registers. GCC on some machines
3595 represents the condition codes as a specific hardware register;
3596 @samp{cc} serves to name this register. On other machines, the
3597 condition code is handled differently, and specifying @samp{cc} has no
3598 effect. But it is valid no matter what the machine.
3600 If your assembler instruction modifies memory in an unpredictable
3601 fashion, add @samp{memory} to the list of clobbered registers. This
3602 will cause GCC to not keep memory values cached in registers across
3603 the assembler instruction. You will also want to add the
3604 @code{volatile} keyword if the memory affected is not listed in the
3605 inputs or outputs of the @code{asm}, as the @samp{memory} clobber does
3606 not count as a side-effect of the @code{asm}.
3608 You can put multiple assembler instructions together in a single
3609 @code{asm} template, separated by the characters normally used in assembly
3610 code for the system. A combination that works in most places is a newline
3611 to break the line, plus a tab character to move to the instruction field
3612 (written as @samp{\n\t}). Sometimes semicolons can be used, if the
3613 assembler allows semicolons as a line-breaking character. Note that some
3614 assembler dialects use semicolons to start a comment.
3615 The input operands are guaranteed not to use any of the clobbered
3616 registers, and neither will the output operands' addresses, so you can
3617 read and write the clobbered registers as many times as you like. Here
3618 is an example of multiple instructions in a template; it assumes the
3619 subroutine @code{_foo} accepts arguments in registers 9 and 10:
3622 asm ("movl %0,r9\n\tmovl %1,r10\n\tcall _foo"
3624 : "g" (from), "g" (to)
3628 Unless an output operand has the @samp{&} constraint modifier, GCC
3629 may allocate it in the same register as an unrelated input operand, on
3630 the assumption the inputs are consumed before the outputs are produced.
3631 This assumption may be false if the assembler code actually consists of
3632 more than one instruction. In such a case, use @samp{&} for each output
3633 operand that may not overlap an input. @xref{Modifiers}.
3635 If you want to test the condition code produced by an assembler
3636 instruction, you must include a branch and a label in the @code{asm}
3637 construct, as follows:
3640 asm ("clr %0\n\tfrob %1\n\tbeq 0f\n\tmov #1,%0\n0:"
3646 This assumes your assembler supports local labels, as the GNU assembler
3647 and most Unix assemblers do.
3649 Speaking of labels, jumps from one @code{asm} to another are not
3650 supported. The compiler's optimizers do not know about these jumps, and
3651 therefore they cannot take account of them when deciding how to
3654 @cindex macros containing @code{asm}
3655 Usually the most convenient way to use these @code{asm} instructions is to
3656 encapsulate them in macros that look like functions. For example,
3660 (@{ double __value, __arg = (x); \
3661 asm ("fsinx %1,%0": "=f" (__value): "f" (__arg)); \
3666 Here the variable @code{__arg} is used to make sure that the instruction
3667 operates on a proper @code{double} value, and to accept only those
3668 arguments @code{x} which can convert automatically to a @code{double}.
3670 Another way to make sure the instruction operates on the correct data
3671 type is to use a cast in the @code{asm}. This is different from using a
3672 variable @code{__arg} in that it converts more different types. For
3673 example, if the desired type were @code{int}, casting the argument to
3674 @code{int} would accept a pointer with no complaint, while assigning the
3675 argument to an @code{int} variable named @code{__arg} would warn about
3676 using a pointer unless the caller explicitly casts it.
3678 If an @code{asm} has output operands, GCC assumes for optimization
3679 purposes the instruction has no side effects except to change the output
3680 operands. This does not mean instructions with a side effect cannot be
3681 used, but you must be careful, because the compiler may eliminate them
3682 if the output operands aren't used, or move them out of loops, or
3683 replace two with one if they constitute a common subexpression. Also,
3684 if your instruction does have a side effect on a variable that otherwise
3685 appears not to change, the old value of the variable may be reused later
3686 if it happens to be found in a register.
3688 You can prevent an @code{asm} instruction from being deleted, moved
3689 significantly, or combined, by writing the keyword @code{volatile} after
3690 the @code{asm}. For example:
3693 #define get_and_set_priority(new) \
3695 asm volatile ("get_and_set_priority %0, %1" \
3696 : "=g" (__old) : "g" (new)); \
3701 If you write an @code{asm} instruction with no outputs, GCC will know
3702 the instruction has side-effects and will not delete the instruction or
3703 move it outside of loops.
3705 The @code{volatile} keyword indicates that the instruction has
3706 important side-effects. GCC will not delete a volatile @code{asm} if
3707 it is reachable. (The instruction can still be deleted if GCC can
3708 prove that control-flow will never reach the location of the
3709 instruction.) In addition, GCC will not reschedule instructions
3710 across a volatile @code{asm} instruction. For example:
3713 *(volatile int *)addr = foo;
3714 asm volatile ("eieio" : : );
3718 Assume @code{addr} contains the address of a memory mapped device
3719 register. The PowerPC @code{eieio} instruction (Enforce In-order
3720 Execution of I/O) tells the CPU to make sure that the store to that
3721 device register happens before it issues any other I/O@.
3723 Note that even a volatile @code{asm} instruction can be moved in ways
3724 that appear insignificant to the compiler, such as across jump
3725 instructions. You can't expect a sequence of volatile @code{asm}
3726 instructions to remain perfectly consecutive. If you want consecutive
3727 output, use a single @code{asm}. Also, GCC will perform some
3728 optimizations across a volatile @code{asm} instruction; GCC does not
3729 ``forget everything'' when it encounters a volatile @code{asm}
3730 instruction the way some other compilers do.
3732 An @code{asm} instruction without any operands or clobbers (an ``old
3733 style'' @code{asm}) will be treated identically to a volatile
3734 @code{asm} instruction.
3736 It is a natural idea to look for a way to give access to the condition
3737 code left by the assembler instruction. However, when we attempted to
3738 implement this, we found no way to make it work reliably. The problem
3739 is that output operands might need reloading, which would result in
3740 additional following ``store'' instructions. On most machines, these
3741 instructions would alter the condition code before there was time to
3742 test it. This problem doesn't arise for ordinary ``test'' and
3743 ``compare'' instructions because they don't have any output operands.
3745 For reasons similar to those described above, it is not possible to give
3746 an assembler instruction access to the condition code left by previous
3749 If you are writing a header file that should be includable in ISO C
3750 programs, write @code{__asm__} instead of @code{asm}. @xref{Alternate
3753 @subsection i386 floating point asm operands
3755 There are several rules on the usage of stack-like regs in
3756 asm_operands insns. These rules apply only to the operands that are
3761 Given a set of input regs that die in an asm_operands, it is
3762 necessary to know which are implicitly popped by the asm, and
3763 which must be explicitly popped by gcc.
3765 An input reg that is implicitly popped by the asm must be
3766 explicitly clobbered, unless it is constrained to match an
3770 For any input reg that is implicitly popped by an asm, it is
3771 necessary to know how to adjust the stack to compensate for the pop.
3772 If any non-popped input is closer to the top of the reg-stack than
3773 the implicitly popped reg, it would not be possible to know what the
3774 stack looked like---it's not clear how the rest of the stack ``slides
3777 All implicitly popped input regs must be closer to the top of
3778 the reg-stack than any input that is not implicitly popped.
3780 It is possible that if an input dies in an insn, reload might
3781 use the input reg for an output reload. Consider this example:
3784 asm ("foo" : "=t" (a) : "f" (b));
3787 This asm says that input B is not popped by the asm, and that
3788 the asm pushes a result onto the reg-stack, i.e., the stack is one
3789 deeper after the asm than it was before. But, it is possible that
3790 reload will think that it can use the same reg for both the input and
3791 the output, if input B dies in this insn.
3793 If any input operand uses the @code{f} constraint, all output reg
3794 constraints must use the @code{&} earlyclobber.
3796 The asm above would be written as
3799 asm ("foo" : "=&t" (a) : "f" (b));
3803 Some operands need to be in particular places on the stack. All
3804 output operands fall in this category---there is no other way to
3805 know which regs the outputs appear in unless the user indicates
3806 this in the constraints.
3808 Output operands must specifically indicate which reg an output
3809 appears in after an asm. @code{=f} is not allowed: the operand
3810 constraints must select a class with a single reg.
3813 Output operands may not be ``inserted'' between existing stack regs.
3814 Since no 387 opcode uses a read/write operand, all output operands
3815 are dead before the asm_operands, and are pushed by the asm_operands.
3816 It makes no sense to push anywhere but the top of the reg-stack.
3818 Output operands must start at the top of the reg-stack: output
3819 operands may not ``skip'' a reg.
3822 Some asm statements may need extra stack space for internal
3823 calculations. This can be guaranteed by clobbering stack registers
3824 unrelated to the inputs and outputs.
3828 Here are a couple of reasonable asms to want to write. This asm
3829 takes one input, which is internally popped, and produces two outputs.
3832 asm ("fsincos" : "=t" (cos), "=u" (sin) : "0" (inp));
3835 This asm takes two inputs, which are popped by the @code{fyl2xp1} opcode,
3836 and replaces them with one output. The user must code the @code{st(1)}
3837 clobber for reg-stack.c to know that @code{fyl2xp1} pops both inputs.
3840 asm ("fyl2xp1" : "=t" (result) : "0" (x), "u" (y) : "st(1)");
3846 @section Controlling Names Used in Assembler Code
3847 @cindex assembler names for identifiers
3848 @cindex names used in assembler code
3849 @cindex identifiers, names in assembler code
3851 You can specify the name to be used in the assembler code for a C
3852 function or variable by writing the @code{asm} (or @code{__asm__})
3853 keyword after the declarator as follows:
3856 int foo asm ("myfoo") = 2;
3860 This specifies that the name to be used for the variable @code{foo} in
3861 the assembler code should be @samp{myfoo} rather than the usual
3864 On systems where an underscore is normally prepended to the name of a C
3865 function or variable, this feature allows you to define names for the
3866 linker that do not start with an underscore.
3868 It does not make sense to use this feature with a non-static local
3869 variable since such variables do not have assembler names. If you are
3870 trying to put the variable in a particular register, see @ref{Explicit
3871 Reg Vars}. GCC presently accepts such code with a warning, but will
3872 probably be changed to issue an error, rather than a warning, in the
3875 You cannot use @code{asm} in this way in a function @emph{definition}; but
3876 you can get the same effect by writing a declaration for the function
3877 before its definition and putting @code{asm} there, like this:
3880 extern func () asm ("FUNC");
3887 It is up to you to make sure that the assembler names you choose do not
3888 conflict with any other assembler symbols. Also, you must not use a
3889 register name; that would produce completely invalid assembler code. GCC
3890 does not as yet have the ability to store static variables in registers.
3891 Perhaps that will be added.
3893 @node Explicit Reg Vars
3894 @section Variables in Specified Registers
3895 @cindex explicit register variables
3896 @cindex variables in specified registers
3897 @cindex specified registers
3898 @cindex registers, global allocation
3900 GNU C allows you to put a few global variables into specified hardware
3901 registers. You can also specify the register in which an ordinary
3902 register variable should be allocated.
3906 Global register variables reserve registers throughout the program.
3907 This may be useful in programs such as programming language
3908 interpreters which have a couple of global variables that are accessed
3912 Local register variables in specific registers do not reserve the
3913 registers. The compiler's data flow analysis is capable of determining
3914 where the specified registers contain live values, and where they are
3915 available for other uses. Stores into local register variables may be deleted
3916 when they appear to be dead according to dataflow analysis. References
3917 to local register variables may be deleted or moved or simplified.
3919 These local variables are sometimes convenient for use with the extended
3920 @code{asm} feature (@pxref{Extended Asm}), if you want to write one
3921 output of the assembler instruction directly into a particular register.
3922 (This will work provided the register you specify fits the constraints
3923 specified for that operand in the @code{asm}.)
3931 @node Global Reg Vars
3932 @subsection Defining Global Register Variables
3933 @cindex global register variables
3934 @cindex registers, global variables in
3936 You can define a global register variable in GNU C like this:
3939 register int *foo asm ("a5");
3943 Here @code{a5} is the name of the register which should be used. Choose a
3944 register which is normally saved and restored by function calls on your
3945 machine, so that library routines will not clobber it.
3947 Naturally the register name is cpu-dependent, so you would need to
3948 conditionalize your program according to cpu type. The register
3949 @code{a5} would be a good choice on a 68000 for a variable of pointer
3950 type. On machines with register windows, be sure to choose a ``global''
3951 register that is not affected magically by the function call mechanism.
3953 In addition, operating systems on one type of cpu may differ in how they
3954 name the registers; then you would need additional conditionals. For
3955 example, some 68000 operating systems call this register @code{%a5}.
3957 Eventually there may be a way of asking the compiler to choose a register
3958 automatically, but first we need to figure out how it should choose and
3959 how to enable you to guide the choice. No solution is evident.
3961 Defining a global register variable in a certain register reserves that
3962 register entirely for this use, at least within the current compilation.
3963 The register will not be allocated for any other purpose in the functions
3964 in the current compilation. The register will not be saved and restored by
3965 these functions. Stores into this register are never deleted even if they
3966 would appear to be dead, but references may be deleted or moved or
3969 It is not safe to access the global register variables from signal
3970 handlers, or from more than one thread of control, because the system
3971 library routines may temporarily use the register for other things (unless
3972 you recompile them specially for the task at hand).
3974 @cindex @code{qsort}, and global register variables
3975 It is not safe for one function that uses a global register variable to
3976 call another such function @code{foo} by way of a third function
3977 @code{lose} that was compiled without knowledge of this variable (i.e.@: in a
3978 different source file in which the variable wasn't declared). This is
3979 because @code{lose} might save the register and put some other value there.
3980 For example, you can't expect a global register variable to be available in
3981 the comparison-function that you pass to @code{qsort}, since @code{qsort}
3982 might have put something else in that register. (If you are prepared to
3983 recompile @code{qsort} with the same global register variable, you can
3984 solve this problem.)
3986 If you want to recompile @code{qsort} or other source files which do not
3987 actually use your global register variable, so that they will not use that
3988 register for any other purpose, then it suffices to specify the compiler
3989 option @option{-ffixed-@var{reg}}. You need not actually add a global
3990 register declaration to their source code.
3992 A function which can alter the value of a global register variable cannot
3993 safely be called from a function compiled without this variable, because it
3994 could clobber the value the caller expects to find there on return.
3995 Therefore, the function which is the entry point into the part of the
3996 program that uses the global register variable must explicitly save and
3997 restore the value which belongs to its caller.
3999 @cindex register variable after @code{longjmp}
4000 @cindex global register after @code{longjmp}
4001 @cindex value after @code{longjmp}
4004 On most machines, @code{longjmp} will restore to each global register
4005 variable the value it had at the time of the @code{setjmp}. On some
4006 machines, however, @code{longjmp} will not change the value of global
4007 register variables. To be portable, the function that called @code{setjmp}
4008 should make other arrangements to save the values of the global register
4009 variables, and to restore them in a @code{longjmp}. This way, the same
4010 thing will happen regardless of what @code{longjmp} does.
4012 All global register variable declarations must precede all function
4013 definitions. If such a declaration could appear after function
4014 definitions, the declaration would be too late to prevent the register from
4015 being used for other purposes in the preceding functions.
4017 Global register variables may not have initial values, because an
4018 executable file has no means to supply initial contents for a register.
4020 On the Sparc, there are reports that g3 @dots{} g7 are suitable
4021 registers, but certain library functions, such as @code{getwd}, as well
4022 as the subroutines for division and remainder, modify g3 and g4. g1 and
4023 g2 are local temporaries.
4025 On the 68000, a2 @dots{} a5 should be suitable, as should d2 @dots{} d7.
4026 Of course, it will not do to use more than a few of those.
4028 @node Local Reg Vars
4029 @subsection Specifying Registers for Local Variables
4030 @cindex local variables, specifying registers
4031 @cindex specifying registers for local variables
4032 @cindex registers for local variables
4034 You can define a local register variable with a specified register
4038 register int *foo asm ("a5");
4042 Here @code{a5} is the name of the register which should be used. Note
4043 that this is the same syntax used for defining global register
4044 variables, but for a local variable it would appear within a function.
4046 Naturally the register name is cpu-dependent, but this is not a
4047 problem, since specific registers are most often useful with explicit
4048 assembler instructions (@pxref{Extended Asm}). Both of these things
4049 generally require that you conditionalize your program according to
4052 In addition, operating systems on one type of cpu may differ in how they
4053 name the registers; then you would need additional conditionals. For
4054 example, some 68000 operating systems call this register @code{%a5}.
4056 Defining such a register variable does not reserve the register; it
4057 remains available for other uses in places where flow control determines
4058 the variable's value is not live. However, these registers are made
4059 unavailable for use in the reload pass; excessive use of this feature
4060 leaves the compiler too few available registers to compile certain
4063 This option does not guarantee that GCC will generate code that has
4064 this variable in the register you specify at all times. You may not
4065 code an explicit reference to this register in an @code{asm} statement
4066 and assume it will always refer to this variable.
4068 Stores into local register variables may be deleted when they appear to be dead
4069 according to dataflow analysis. References to local register variables may
4070 be deleted or moved or simplified.
4072 @node Alternate Keywords
4073 @section Alternate Keywords
4074 @cindex alternate keywords
4075 @cindex keywords, alternate
4077 @option{-ansi} and the various @option{-std} options disable certain
4078 keywords. This causes trouble when you want to use GNU C extensions, or
4079 a general-purpose header file that should be usable by all programs,
4080 including ISO C programs. The keywords @code{asm}, @code{typeof} and
4081 @code{inline} are not available in programs compiled with
4082 @option{-ansi} or @option{-std} (although @code{inline} can be used in a
4083 program compiled with @option{-std=c99}). The ISO C99 keyword
4084 @code{restrict} is only available when @option{-std=gnu99} (which will
4085 eventually be the default) or @option{-std=c99} (or the equivalent
4086 @option{-std=iso9899:1999}) is used.
4088 The way to solve these problems is to put @samp{__} at the beginning and
4089 end of each problematical keyword. For example, use @code{__asm__}
4090 instead of @code{asm}, and @code{__inline__} instead of @code{inline}.
4092 Other C compilers won't accept these alternative keywords; if you want to
4093 compile with another compiler, you can define the alternate keywords as
4094 macros to replace them with the customary keywords. It looks like this:
4102 @findex __extension__
4104 @option{-pedantic} and other options cause warnings for many GNU C extensions.
4106 prevent such warnings within one expression by writing
4107 @code{__extension__} before the expression. @code{__extension__} has no
4108 effect aside from this.
4110 @node Incomplete Enums
4111 @section Incomplete @code{enum} Types
4113 You can define an @code{enum} tag without specifying its possible values.
4114 This results in an incomplete type, much like what you get if you write
4115 @code{struct foo} without describing the elements. A later declaration
4116 which does specify the possible values completes the type.
4118 You can't allocate variables or storage using the type while it is
4119 incomplete. However, you can work with pointers to that type.
4121 This extension may not be very useful, but it makes the handling of
4122 @code{enum} more consistent with the way @code{struct} and @code{union}
4125 This extension is not supported by GNU C++.
4127 @node Function Names
4128 @section Function Names as Strings
4129 @cindex @code{__FUNCTION__} identifier
4130 @cindex @code{__PRETTY_FUNCTION__} identifier
4131 @cindex @code{__func__} identifier
4133 GCC predefines two magic identifiers to hold the name of the current
4134 function. The identifier @code{__FUNCTION__} holds the name of the function
4135 as it appears in the source. The identifier @code{__PRETTY_FUNCTION__}
4136 holds the name of the function pretty printed in a language specific
4139 These names are always the same in a C function, but in a C++ function
4140 they may be different. For example, this program:
4144 extern int printf (char *, ...);
4151 printf ("__FUNCTION__ = %s\n", __FUNCTION__);
4152 printf ("__PRETTY_FUNCTION__ = %s\n", __PRETTY_FUNCTION__);
4170 __PRETTY_FUNCTION__ = int a::sub (int)
4173 The compiler automagically replaces the identifiers with a string
4174 literal containing the appropriate name. Thus, they are neither
4175 preprocessor macros, like @code{__FILE__} and @code{__LINE__}, nor
4176 variables. This means that they catenate with other string literals, and
4177 that they can be used to initialize char arrays. For example
4180 char here[] = "Function " __FUNCTION__ " in " __FILE__;
4183 On the other hand, @samp{#ifdef __FUNCTION__} does not have any special
4184 meaning inside a function, since the preprocessor does not do anything
4185 special with the identifier @code{__FUNCTION__}.
4187 Note that these semantics are deprecated, and that GCC 3.2 will handle
4188 @code{__FUNCTION__} and @code{__PRETTY_FUNCTION__} the same way as
4189 @code{__func__}. @code{__func__} is defined by the ISO standard C99:
4192 The identifier @code{__func__} is implicitly declared by the translator
4193 as if, immediately following the opening brace of each function
4194 definition, the declaration
4197 static const char __func__[] = "function-name";
4200 appeared, where function-name is the name of the lexically-enclosing
4201 function. This name is the unadorned name of the function.
4204 By this definition, @code{__func__} is a variable, not a string literal.
4205 In particular, @code{__func__} does not catenate with other string
4208 In @code{C++}, @code{__FUNCTION__} and @code{__PRETTY_FUNCTION__} are
4209 variables, declared in the same way as @code{__func__}.
4211 @node Return Address
4212 @section Getting the Return or Frame Address of a Function
4214 These functions may be used to get information about the callers of a
4217 @deftypefn {Built-in Function} {void *} __builtin_return_address (unsigned int @var{level})
4218 This function returns the return address of the current function, or of
4219 one of its callers. The @var{level} argument is number of frames to
4220 scan up the call stack. A value of @code{0} yields the return address
4221 of the current function, a value of @code{1} yields the return address
4222 of the caller of the current function, and so forth.
4224 The @var{level} argument must be a constant integer.
4226 On some machines it may be impossible to determine the return address of
4227 any function other than the current one; in such cases, or when the top
4228 of the stack has been reached, this function will return @code{0} or a
4229 random value. In addition, @code{__builtin_frame_address} may be used
4230 to determine if the top of the stack has been reached.
4232 This function should only be used with a nonzero argument for debugging
4236 @deftypefn {Built-in Function} {void *} __builtin_frame_address (unsigned int @var{level})
4237 This function is similar to @code{__builtin_return_address}, but it
4238 returns the address of the function frame rather than the return address
4239 of the function. Calling @code{__builtin_frame_address} with a value of
4240 @code{0} yields the frame address of the current function, a value of
4241 @code{1} yields the frame address of the caller of the current function,
4244 The frame is the area on the stack which holds local variables and saved
4245 registers. The frame address is normally the address of the first word
4246 pushed on to the stack by the function. However, the exact definition
4247 depends upon the processor and the calling convention. If the processor
4248 has a dedicated frame pointer register, and the function has a frame,
4249 then @code{__builtin_frame_address} will return the value of the frame
4252 On some machines it may be impossible to determine the frame address of
4253 any function other than the current one; in such cases, or when the top
4254 of the stack has been reached, this function will return @code{0} if
4255 the first frame pointer is properly initialized by the startup code.
4257 This function should only be used with a nonzero argument for debugging
4261 @node Vector Extensions
4262 @section Using vector instructions through built-in functions
4264 On some targets, the instruction set contains SIMD vector instructions that
4265 operate on multiple values contained in one large register at the same time.
4266 For example, on the i386 the MMX, 3Dnow! and SSE extensions can be used
4269 The first step in using these extensions is to provide the necessary data
4270 types. This should be done using an appropriate @code{typedef}:
4273 typedef int v4si __attribute__ ((mode(V4SI)));
4276 The base type @code{int} is effectively ignored by the compiler, the
4277 actual properties of the new type @code{v4si} are defined by the
4278 @code{__attribute__}. It defines the machine mode to be used; for vector
4279 types these have the form @code{V@var{n}@var{B}}; @var{n} should be the
4280 number of elements in the vector, and @var{B} should be the base mode of the
4281 individual elements. The following can be used as base modes:
4285 An integer that is as wide as the smallest addressable unit, usually 8 bits.
4287 An integer, twice as wide as a QI mode integer, usually 16 bits.
4289 An integer, four times as wide as a QI mode integer, usually 32 bits.
4291 An integer, eight times as wide as a QI mode integer, usually 64 bits.
4293 A floating point value, as wide as a SI mode integer, usually 32 bits.
4295 A floating point value, as wide as a DI mode integer, usually 64 bits.
4298 Not all base types or combinations are always valid; which modes can be used
4299 is determined by the target machine. For example, if targetting the i386 MMX
4300 extensions, only @code{V8QI}, @code{V4HI} and @code{V2SI} are allowed modes.
4302 There are no @code{V1xx} vector modes - they would be identical to the
4303 corresponding base mode.
4305 There is no distinction between signed and unsigned vector modes. This
4306 distinction is made by the operations that perform on the vectors, not
4309 The types defined in this manner are somewhat special, they cannot be
4310 used with most normal C operations (i.e., a vector addition can @emph{not}
4311 be represented by a normal addition of two vector type variables). You
4312 can declare only variables and use them in function calls and returns, as
4313 well as in assignments and some casts. It is possible to cast from one
4314 vector type to another, provided they are of the same size (in fact, you
4315 can also cast vectors to and from other datatypes of the same size).
4317 A port that supports vector operations provides a set of built-in functions
4318 that can be used to operate on vectors. For example, a function to add two
4319 vectors and multiply the result by a third could look like this:
4322 v4si f (v4si a, v4si b, v4si c)
4324 v4si tmp = __builtin_addv4si (a, b);
4325 return __builtin_mulv4si (tmp, c);
4330 @node Other Builtins
4331 @section Other built-in functions provided by GCC
4332 @cindex built-in functions
4333 @findex __builtin_isgreater
4334 @findex __builtin_isgreaterequal
4335 @findex __builtin_isless
4336 @findex __builtin_islessequal
4337 @findex __builtin_islessgreater
4338 @findex __builtin_isunordered
4364 @findex fprintf_unlocked
4366 @findex fputs_unlocked
4375 @findex printf_unlocked
4397 GCC provides a large number of built-in functions other than the ones
4398 mentioned above. Some of these are for internal use in the processing
4399 of exceptions or variable-length argument lists and will not be
4400 documented here because they may change from time to time; we do not
4401 recommend general use of these functions.
4403 The remaining functions are provided for optimization purposes.
4405 @opindex fno-builtin
4406 GCC includes built-in versions of many of the functions in the standard
4407 C library. The versions prefixed with @code{__builtin_} will always be
4408 treated as having the same meaning as the C library function even if you
4409 specify the @option{-fno-builtin} option. (@pxref{C Dialect Options})
4410 Many of these functions are only optimized in certain cases; if they are
4411 not optimized in a particular case, a call to the library function will
4416 The functions @code{abort}, @code{exit}, @code{_Exit} and @code{_exit}
4417 are recognized and presumed not to return, but otherwise are not built
4418 in. @code{_exit} is not recognized in strict ISO C mode (@option{-ansi},
4419 @option{-std=c89} or @option{-std=c99}). @code{_Exit} is not recognized in
4420 strict C89 mode (@option{-ansi} or @option{-std=c89}).
4422 Outside strict ISO C mode, the functions @code{alloca}, @code{bcmp},
4423 @code{bzero}, @code{index}, @code{rindex}, @code{ffs}, @code{fputs_unlocked},
4424 @code{printf_unlocked} and @code{fprintf_unlocked} may be handled as
4425 built-in functions. All these functions have corresponding versions
4426 prefixed with @code{__builtin_}, which may be used even in strict C89
4429 The ISO C99 functions @code{conj}, @code{conjf}, @code{conjl},
4430 @code{creal}, @code{crealf}, @code{creall}, @code{cimag}, @code{cimagf},
4431 @code{cimagl}, @code{llabs} and @code{imaxabs} are handled as built-in
4432 functions except in strict ISO C89 mode. There are also built-in
4433 versions of the ISO C99 functions @code{cosf}, @code{cosl},
4434 @code{fabsf}, @code{fabsl}, @code{sinf}, @code{sinl}, @code{sqrtf}, and
4435 @code{sqrtl}, that are recognized in any mode since ISO C89 reserves
4436 these names for the purpose to which ISO C99 puts them. All these
4437 functions have corresponding versions prefixed with @code{__builtin_}.
4439 The ISO C89 functions @code{abs}, @code{cos}, @code{fabs},
4440 @code{fprintf}, @code{fputs}, @code{labs}, @code{memcmp}, @code{memcpy},
4441 @code{memset}, @code{printf}, @code{sin}, @code{sqrt}, @code{strcat},
4442 @code{strchr}, @code{strcmp}, @code{strcpy}, @code{strcspn},
4443 @code{strlen}, @code{strncat}, @code{strncmp}, @code{strncpy},
4444 @code{strpbrk}, @code{strrchr}, @code{strspn}, and @code{strstr} are all
4445 recognized as built-in functions unless @option{-fno-builtin} is
4446 specified (or @option{-fno-builtin-@var{function}} is specified for an
4447 individual function). All of these functions have corresponding
4448 versions prefixed with @code{__builtin_}.
4450 GCC provides built-in versions of the ISO C99 floating point comparison
4451 macros that avoid raising exceptions for unordered operands. They have
4452 the same names as the standard macros ( @code{isgreater},
4453 @code{isgreaterequal}, @code{isless}, @code{islessequal},
4454 @code{islessgreater}, and @code{isunordered}) , with @code{__builtin_}
4455 prefixed. We intend for a library implementor to be able to simply
4456 @code{#define} each standard macro to its built-in equivalent.
4458 @deftypefn {Built-in Function} int __builtin_types_compatible_p (@var{type1}, @var{type2})
4460 You can use the built-in function @code{__builtin_types_compatible_p} to
4461 determine whether two types are the same.
4463 This built-in function returns 1 if the unqualified versions of the
4464 types @var{type1} and @var{type2} (which are types, not expressions) are
4465 compatible, 0 otherwise. The result of this built-in function can be
4466 used in integer constant expressions.
4468 This built-in function ignores top level qualifiers (e.g., @code{const},
4469 @code{volatile}). For example, @code{int} is equivalent to @code{const
4472 The type @code{int[]} and @code{int[5]} are compatible. On the other
4473 hand, @code{int} and @code{char *} are not compatible, even if the size
4474 of their types, on the particular architecture are the same. Also, the
4475 amount of pointer indirection is taken into account when determining
4476 similarity. Consequently, @code{short *} is not similar to
4477 @code{short **}. Furthermore, two types that are typedefed are
4478 considered compatible if their underlying types are compatible.
4480 An @code{enum} type is considered to be compatible with another
4481 @code{enum} type. For example, @code{enum @{foo, bar@}} is similar to
4482 @code{enum @{hot, dog@}}.
4484 You would typically use this function in code whose execution varies
4485 depending on the arguments' types. For example:
4491 if (__builtin_types_compatible_p (typeof (x), long double)) \
4492 tmp = foo_long_double (tmp); \
4493 else if (__builtin_types_compatible_p (typeof (x), double)) \
4494 tmp = foo_double (tmp); \
4495 else if (__builtin_types_compatible_p (typeof (x), float)) \
4496 tmp = foo_float (tmp); \
4503 @emph{Note:} This construct is only available for C.
4507 @deftypefn {Built-in Function} @var{type} __builtin_choose_expr (@var{const_exp}, @var{exp1}, @var{exp2})
4509 You can use the built-in function @code{__builtin_choose_expr} to
4510 evaluate code depending on the value of a constant expression. This
4511 built-in function returns @var{exp1} if @var{const_exp}, which is a
4512 constant expression that must be able to be determined at compile time,
4513 is nonzero. Otherwise it returns 0.
4515 This built-in function is analogous to the @samp{? :} operator in C,
4516 except that the expression returned has its type unaltered by promotion
4517 rules. Also, the built-in function does not evaluate the expression
4518 that was not chosen. For example, if @var{const_exp} evaluates to true,
4519 @var{exp2} is not evaluated even if it has side-effects.
4521 This built-in function can return an lvalue if the chosen argument is an
4524 If @var{exp1} is returned, the return type is the same as @var{exp1}'s
4525 type. Similarly, if @var{exp2} is returned, its return type is the same
4532 __builtin_choose_expr (__builtin_types_compatible_p (typeof (x), double), \
4534 __builtin_choose_expr (__builtin_types_compatible_p (typeof (x), float), \
4536 /* @r{The void expression results in a compile-time error} \
4537 @r{when assigning the result to something.} */ \
4541 @emph{Note:} This construct is only available for C. Furthermore, the
4542 unused expression (@var{exp1} or @var{exp2} depending on the value of
4543 @var{const_exp}) may still generate syntax errors. This may change in
4548 @deftypefn {Built-in Function} int __builtin_constant_p (@var{exp})
4549 You can use the built-in function @code{__builtin_constant_p} to
4550 determine if a value is known to be constant at compile-time and hence
4551 that GCC can perform constant-folding on expressions involving that
4552 value. The argument of the function is the value to test. The function
4553 returns the integer 1 if the argument is known to be a compile-time
4554 constant and 0 if it is not known to be a compile-time constant. A
4555 return of 0 does not indicate that the value is @emph{not} a constant,
4556 but merely that GCC cannot prove it is a constant with the specified
4557 value of the @option{-O} option.
4559 You would typically use this function in an embedded application where
4560 memory was a critical resource. If you have some complex calculation,
4561 you may want it to be folded if it involves constants, but need to call
4562 a function if it does not. For example:
4565 #define Scale_Value(X) \
4566 (__builtin_constant_p (X) \
4567 ? ((X) * SCALE + OFFSET) : Scale (X))
4570 You may use this built-in function in either a macro or an inline
4571 function. However, if you use it in an inlined function and pass an
4572 argument of the function as the argument to the built-in, GCC will
4573 never return 1 when you call the inline function with a string constant
4574 or compound literal (@pxref{Compound Literals}) and will not return 1
4575 when you pass a constant numeric value to the inline function unless you
4576 specify the @option{-O} option.
4578 You may also use @code{__builtin_constant_p} in initializers for static
4579 data. For instance, you can write
4582 static const int table[] = @{
4583 __builtin_constant_p (EXPRESSION) ? (EXPRESSION) : -1,
4589 This is an acceptable initializer even if @var{EXPRESSION} is not a
4590 constant expression. GCC must be more conservative about evaluating the
4591 built-in in this case, because it has no opportunity to perform
4594 Previous versions of GCC did not accept this built-in in data
4595 initializers. The earliest version where it is completely safe is
4599 @deftypefn {Built-in Function} long __builtin_expect (long @var{exp}, long @var{c})
4600 @opindex fprofile-arcs
4601 You may use @code{__builtin_expect} to provide the compiler with
4602 branch prediction information. In general, you should prefer to
4603 use actual profile feedback for this (@option{-fprofile-arcs}), as
4604 programmers are notoriously bad at predicting how their programs
4605 actually perform. However, there are applications in which this
4606 data is hard to collect.
4608 The return value is the value of @var{exp}, which should be an
4609 integral expression. The value of @var{c} must be a compile-time
4610 constant. The semantics of the built-in are that it is expected
4611 that @var{exp} == @var{c}. For example:
4614 if (__builtin_expect (x, 0))
4619 would indicate that we do not expect to call @code{foo}, since
4620 we expect @code{x} to be zero. Since you are limited to integral
4621 expressions for @var{exp}, you should use constructions such as
4624 if (__builtin_expect (ptr != NULL, 1))
4629 when testing pointer or floating-point values.
4632 @deftypefn {Built-in Function} void __builtin_prefetch (const void *@var{addr}, ...)
4633 This function is used to minimize cache-miss latency by moving data into
4634 a cache before it is accessed.
4635 You can insert calls to @code{__builtin_prefetch} into code for which
4636 you know addresses of data in memory that is likely to be accessed soon.
4637 If the target supports them, data prefetch instructions will be generated.
4638 If the prefetch is done early enough before the access then the data will
4639 be in the cache by the time it is accessed.
4641 The value of @var{addr} is the address of the memory to prefetch.
4642 There are two optional arguments, @var{rw} and @var{locality}.
4643 The value of @var{rw} is a compile-time constant one or zero; one
4644 means that the prefetch is preparing for a write to the memory address
4645 and zero, the default, means that the prefetch is preparing for a read.
4646 The value @var{locality} must be a compile-time constant integer between
4647 zero and three. A value of zero means that the data has no temporal
4648 locality, so it need not be left in the cache after the access. A value
4649 of three means that the data has a high degree of temporal locality and
4650 should be left in all levels of cache possible. Values of one and two
4651 mean, respectively, a low or moderate degree of temporal locality. The
4655 for (i = 0; i < n; i++)
4658 __builtin_prefetch (&a[i+j], 1, 1);
4659 __builtin_prefetch (&b[i+j], 0, 1);
4664 Data prefetch does not generate faults if @var{addr} is invalid, but
4665 the address expression itself must be valid. For example, a prefetch
4666 of @code{p->next} will not fault if @code{p->next} is not a valid
4667 address, but evaluation will fault if @code{p} is not a valid address.
4669 If the target does not support data prefetch, the address expression
4670 is evaluated if it includes side effects but no other code is generated
4671 and GCC does not issue a warning.
4674 @node Target Builtins
4675 @section Built-in Functions Specific to Particular Target Machines
4677 On some target machines, GCC supports many built-in functions specific
4678 to those machines. Generally these generate calls to specific machine
4679 instructions, but allow the compiler to schedule those calls.
4682 * X86 Built-in Functions::
4683 * PowerPC AltiVec Built-in Functions::
4686 @node X86 Built-in Functions
4687 @subsection X86 Built-in Functions
4689 These built-in functions are available for the i386 and x86-64 family
4690 of computers, depending on the command-line switches used.
4692 The following machine modes are available for use with MMX built-in functions
4693 (@pxref{Vector Extensions}): @code{V2SI} for a vector of two 32-bit integers,
4694 @code{V4HI} for a vector of four 16-bit integers, and @code{V8QI} for a
4695 vector of eight 8-bit integers. Some of the built-in functions operate on
4696 MMX registers as a whole 64-bit entity, these use @code{DI} as their mode.
4698 If 3Dnow extensions are enabled, @code{V2SF} is used as a mode for a vector
4699 of two 32-bit floating point values.
4701 If SSE extensions are enabled, @code{V4SF} is used for a vector of four 32-bit
4702 floating point values. Some instructions use a vector of four 32-bit
4703 integers, these use @code{V4SI}. Finally, some instructions operate on an
4704 entire vector register, interpreting it as a 128-bit integer, these use mode
4707 The following built-in functions are made available by @option{-mmmx}.
4708 All of them generate the machine instruction that is part of the name.
4711 v8qi __builtin_ia32_paddb (v8qi, v8qi)
4712 v4hi __builtin_ia32_paddw (v4hi, v4hi)
4713 v2si __builtin_ia32_paddd (v2si, v2si)
4714 v8qi __builtin_ia32_psubb (v8qi, v8qi)
4715 v4hi __builtin_ia32_psubw (v4hi, v4hi)
4716 v2si __builtin_ia32_psubd (v2si, v2si)
4717 v8qi __builtin_ia32_paddsb (v8qi, v8qi)
4718 v4hi __builtin_ia32_paddsw (v4hi, v4hi)
4719 v8qi __builtin_ia32_psubsb (v8qi, v8qi)
4720 v4hi __builtin_ia32_psubsw (v4hi, v4hi)
4721 v8qi __builtin_ia32_paddusb (v8qi, v8qi)
4722 v4hi __builtin_ia32_paddusw (v4hi, v4hi)
4723 v8qi __builtin_ia32_psubusb (v8qi, v8qi)
4724 v4hi __builtin_ia32_psubusw (v4hi, v4hi)
4725 v4hi __builtin_ia32_pmullw (v4hi, v4hi)
4726 v4hi __builtin_ia32_pmulhw (v4hi, v4hi)
4727 di __builtin_ia32_pand (di, di)
4728 di __builtin_ia32_pandn (di,di)
4729 di __builtin_ia32_por (di, di)
4730 di __builtin_ia32_pxor (di, di)
4731 v8qi __builtin_ia32_pcmpeqb (v8qi, v8qi)
4732 v4hi __builtin_ia32_pcmpeqw (v4hi, v4hi)
4733 v2si __builtin_ia32_pcmpeqd (v2si, v2si)
4734 v8qi __builtin_ia32_pcmpgtb (v8qi, v8qi)
4735 v4hi __builtin_ia32_pcmpgtw (v4hi, v4hi)
4736 v2si __builtin_ia32_pcmpgtd (v2si, v2si)
4737 v8qi __builtin_ia32_punpckhbw (v8qi, v8qi)
4738 v4hi __builtin_ia32_punpckhwd (v4hi, v4hi)
4739 v2si __builtin_ia32_punpckhdq (v2si, v2si)
4740 v8qi __builtin_ia32_punpcklbw (v8qi, v8qi)
4741 v4hi __builtin_ia32_punpcklwd (v4hi, v4hi)
4742 v2si __builtin_ia32_punpckldq (v2si, v2si)
4743 v8qi __builtin_ia32_packsswb (v4hi, v4hi)
4744 v4hi __builtin_ia32_packssdw (v2si, v2si)
4745 v8qi __builtin_ia32_packuswb (v4hi, v4hi)
4748 The following built-in functions are made available either with
4749 @option{-msse}, or with a combination of @option{-m3dnow} and
4750 @option{-march=athlon}. All of them generate the machine
4751 instruction that is part of the name.
4754 v4hi __builtin_ia32_pmulhuw (v4hi, v4hi)
4755 v8qi __builtin_ia32_pavgb (v8qi, v8qi)
4756 v4hi __builtin_ia32_pavgw (v4hi, v4hi)
4757 v4hi __builtin_ia32_psadbw (v8qi, v8qi)
4758 v8qi __builtin_ia32_pmaxub (v8qi, v8qi)
4759 v4hi __builtin_ia32_pmaxsw (v4hi, v4hi)
4760 v8qi __builtin_ia32_pminub (v8qi, v8qi)
4761 v4hi __builtin_ia32_pminsw (v4hi, v4hi)
4762 int __builtin_ia32_pextrw (v4hi, int)
4763 v4hi __builtin_ia32_pinsrw (v4hi, int, int)
4764 int __builtin_ia32_pmovmskb (v8qi)
4765 void __builtin_ia32_maskmovq (v8qi, v8qi, char *)
4766 void __builtin_ia32_movntq (di *, di)
4767 void __builtin_ia32_sfence (void)
4770 The following built-in functions are available when @option{-msse} is used.
4771 All of them generate the machine instruction that is part of the name.
4774 int __builtin_ia32_comieq (v4sf, v4sf)
4775 int __builtin_ia32_comineq (v4sf, v4sf)
4776 int __builtin_ia32_comilt (v4sf, v4sf)
4777 int __builtin_ia32_comile (v4sf, v4sf)
4778 int __builtin_ia32_comigt (v4sf, v4sf)
4779 int __builtin_ia32_comige (v4sf, v4sf)
4780 int __builtin_ia32_ucomieq (v4sf, v4sf)
4781 int __builtin_ia32_ucomineq (v4sf, v4sf)
4782 int __builtin_ia32_ucomilt (v4sf, v4sf)
4783 int __builtin_ia32_ucomile (v4sf, v4sf)
4784 int __builtin_ia32_ucomigt (v4sf, v4sf)
4785 int __builtin_ia32_ucomige (v4sf, v4sf)
4786 v4sf __builtin_ia32_addps (v4sf, v4sf)
4787 v4sf __builtin_ia32_subps (v4sf, v4sf)
4788 v4sf __builtin_ia32_mulps (v4sf, v4sf)
4789 v4sf __builtin_ia32_divps (v4sf, v4sf)
4790 v4sf __builtin_ia32_addss (v4sf, v4sf)
4791 v4sf __builtin_ia32_subss (v4sf, v4sf)
4792 v4sf __builtin_ia32_mulss (v4sf, v4sf)
4793 v4sf __builtin_ia32_divss (v4sf, v4sf)
4794 v4si __builtin_ia32_cmpeqps (v4sf, v4sf)
4795 v4si __builtin_ia32_cmpltps (v4sf, v4sf)
4796 v4si __builtin_ia32_cmpleps (v4sf, v4sf)
4797 v4si __builtin_ia32_cmpgtps (v4sf, v4sf)
4798 v4si __builtin_ia32_cmpgeps (v4sf, v4sf)
4799 v4si __builtin_ia32_cmpunordps (v4sf, v4sf)
4800 v4si __builtin_ia32_cmpneqps (v4sf, v4sf)
4801 v4si __builtin_ia32_cmpnltps (v4sf, v4sf)
4802 v4si __builtin_ia32_cmpnleps (v4sf, v4sf)
4803 v4si __builtin_ia32_cmpngtps (v4sf, v4sf)
4804 v4si __builtin_ia32_cmpngeps (v4sf, v4sf)
4805 v4si __builtin_ia32_cmpordps (v4sf, v4sf)
4806 v4si __builtin_ia32_cmpeqss (v4sf, v4sf)
4807 v4si __builtin_ia32_cmpltss (v4sf, v4sf)
4808 v4si __builtin_ia32_cmpless (v4sf, v4sf)
4809 v4si __builtin_ia32_cmpgtss (v4sf, v4sf)
4810 v4si __builtin_ia32_cmpgess (v4sf, v4sf)
4811 v4si __builtin_ia32_cmpunordss (v4sf, v4sf)
4812 v4si __builtin_ia32_cmpneqss (v4sf, v4sf)
4813 v4si __builtin_ia32_cmpnlts (v4sf, v4sf)
4814 v4si __builtin_ia32_cmpnless (v4sf, v4sf)
4815 v4si __builtin_ia32_cmpngtss (v4sf, v4sf)
4816 v4si __builtin_ia32_cmpngess (v4sf, v4sf)
4817 v4si __builtin_ia32_cmpordss (v4sf, v4sf)
4818 v4sf __builtin_ia32_maxps (v4sf, v4sf)
4819 v4sf __builtin_ia32_maxss (v4sf, v4sf)
4820 v4sf __builtin_ia32_minps (v4sf, v4sf)
4821 v4sf __builtin_ia32_minss (v4sf, v4sf)
4822 v4sf __builtin_ia32_andps (v4sf, v4sf)
4823 v4sf __builtin_ia32_andnps (v4sf, v4sf)
4824 v4sf __builtin_ia32_orps (v4sf, v4sf)
4825 v4sf __builtin_ia32_xorps (v4sf, v4sf)
4826 v4sf __builtin_ia32_movss (v4sf, v4sf)
4827 v4sf __builtin_ia32_movhlps (v4sf, v4sf)
4828 v4sf __builtin_ia32_movlhps (v4sf, v4sf)
4829 v4sf __builtin_ia32_unpckhps (v4sf, v4sf)
4830 v4sf __builtin_ia32_unpcklps (v4sf, v4sf)
4831 v4sf __builtin_ia32_cvtpi2ps (v4sf, v2si)
4832 v4sf __builtin_ia32_cvtsi2ss (v4sf, int)
4833 v2si __builtin_ia32_cvtps2pi (v4sf)
4834 int __builtin_ia32_cvtss2si (v4sf)
4835 v2si __builtin_ia32_cvttps2pi (v4sf)
4836 int __builtin_ia32_cvttss2si (v4sf)
4837 v4sf __builtin_ia32_rcpps (v4sf)
4838 v4sf __builtin_ia32_rsqrtps (v4sf)
4839 v4sf __builtin_ia32_sqrtps (v4sf)
4840 v4sf __builtin_ia32_rcpss (v4sf)
4841 v4sf __builtin_ia32_rsqrtss (v4sf)
4842 v4sf __builtin_ia32_sqrtss (v4sf)
4843 v4sf __builtin_ia32_shufps (v4sf, v4sf, int)
4844 void __builtin_ia32_movntps (float *, v4sf)
4845 int __builtin_ia32_movmskps (v4sf)
4848 The following built-in functions are available when @option{-msse} is used.
4851 @item v4sf __builtin_ia32_loadaps (float *)
4852 Generates the @code{movaps} machine instruction as a load from memory.
4853 @item void __builtin_ia32_storeaps (float *, v4sf)
4854 Generates the @code{movaps} machine instruction as a store to memory.
4855 @item v4sf __builtin_ia32_loadups (float *)
4856 Generates the @code{movups} machine instruction as a load from memory.
4857 @item void __builtin_ia32_storeups (float *, v4sf)
4858 Generates the @code{movups} machine instruction as a store to memory.
4859 @item v4sf __builtin_ia32_loadsss (float *)
4860 Generates the @code{movss} machine instruction as a load from memory.
4861 @item void __builtin_ia32_storess (float *, v4sf)
4862 Generates the @code{movss} machine instruction as a store to memory.
4863 @item v4sf __builtin_ia32_loadhps (v4sf, v2si *)
4864 Generates the @code{movhps} machine instruction as a load from memory.
4865 @item v4sf __builtin_ia32_loadlps (v4sf, v2si *)
4866 Generates the @code{movlps} machine instruction as a load from memory
4867 @item void __builtin_ia32_storehps (v4sf, v2si *)
4868 Generates the @code{movhps} machine instruction as a store to memory.
4869 @item void __builtin_ia32_storelps (v4sf, v2si *)
4870 Generates the @code{movlps} machine instruction as a store to memory.
4873 The following built-in functions are available when @option{-m3dnow} is used.
4874 All of them generate the machine instruction that is part of the name.
4877 void __builtin_ia32_femms (void)
4878 v8qi __builtin_ia32_pavgusb (v8qi, v8qi)
4879 v2si __builtin_ia32_pf2id (v2sf)
4880 v2sf __builtin_ia32_pfacc (v2sf, v2sf)
4881 v2sf __builtin_ia32_pfadd (v2sf, v2sf)
4882 v2si __builtin_ia32_pfcmpeq (v2sf, v2sf)
4883 v2si __builtin_ia32_pfcmpge (v2sf, v2sf)
4884 v2si __builtin_ia32_pfcmpgt (v2sf, v2sf)
4885 v2sf __builtin_ia32_pfmax (v2sf, v2sf)
4886 v2sf __builtin_ia32_pfmin (v2sf, v2sf)
4887 v2sf __builtin_ia32_pfmul (v2sf, v2sf)
4888 v2sf __builtin_ia32_pfrcp (v2sf)
4889 v2sf __builtin_ia32_pfrcpit1 (v2sf, v2sf)
4890 v2sf __builtin_ia32_pfrcpit2 (v2sf, v2sf)
4891 v2sf __builtin_ia32_pfrsqrt (v2sf)
4892 v2sf __builtin_ia32_pfrsqrtit1 (v2sf, v2sf)
4893 v2sf __builtin_ia32_pfsub (v2sf, v2sf)
4894 v2sf __builtin_ia32_pfsubr (v2sf, v2sf)
4895 v2sf __builtin_ia32_pi2fd (v2si)
4896 v4hi __builtin_ia32_pmulhrw (v4hi, v4hi)
4899 The following built-in functions are available when both @option{-m3dnow}
4900 and @option{-march=athlon} are used. All of them generate the machine
4901 instruction that is part of the name.
4904 v2si __builtin_ia32_pf2iw (v2sf)
4905 v2sf __builtin_ia32_pfnacc (v2sf, v2sf)
4906 v2sf __builtin_ia32_pfpnacc (v2sf, v2sf)
4907 v2sf __builtin_ia32_pi2fw (v2si)
4908 v2sf __builtin_ia32_pswapdsf (v2sf)
4909 v2si __builtin_ia32_pswapdsi (v2si)
4912 @node PowerPC AltiVec Built-in Functions
4913 @subsection PowerPC AltiVec Built-in Functions
4915 These built-in functions are available for the PowerPC family
4916 of computers, depending on the command-line switches used.
4918 The following machine modes are available for use with AltiVec built-in
4919 functions (@pxref{Vector Extensions}): @code{V4SI} for a vector of four
4920 32-bit integers, @code{V4SF} for a vector of four 32-bit floating point
4921 numbers, @code{V8HI} for a vector of eight 16-bit integers, and
4922 @code{V16QI} for a vector of sixteen 8-bit integers.
4924 The following functions are made available by including
4925 @code{<altivec.h>} and using @option{-maltivec} and
4926 @option{-mabi=altivec}. The functions implement the functionality
4927 described in Motorola's AltiVec Programming Interface Manual.
4929 @emph{Note:} Only the @code{<altivec.h>} interface is supported.
4930 Internally, GCC uses built-in functions to achieve the functionality in
4931 the aforementioned header file, but they are not supported and are
4932 subject to change without notice.
4935 vector signed char vec_abs (vector signed char, vector signed char);
4936 vector signed short vec_abs (vector signed short, vector signed short);
4937 vector signed int vec_abs (vector signed int, vector signed int);
4938 vector signed float vec_abs (vector signed float, vector signed float);
4940 vector signed char vec_abss (vector signed char, vector signed char);
4941 vector signed short vec_abss (vector signed short, vector signed short);
4943 vector signed char vec_add (vector signed char, vector signed char);
4944 vector unsigned char vec_add (vector signed char, vector unsigned char);
4946 vector unsigned char vec_add (vector unsigned char, vector signed char);
4948 vector unsigned char vec_add (vector unsigned char,
4949 vector unsigned char);
4950 vector signed short vec_add (vector signed short, vector signed short);
4951 vector unsigned short vec_add (vector signed short,
4952 vector unsigned short);
4953 vector unsigned short vec_add (vector unsigned short,
4954 vector signed short);
4955 vector unsigned short vec_add (vector unsigned short,
4956 vector unsigned short);
4957 vector signed int vec_add (vector signed int, vector signed int);
4958 vector unsigned int vec_add (vector signed int, vector unsigned int);
4959 vector unsigned int vec_add (vector unsigned int, vector signed int);
4960 vector unsigned int vec_add (vector unsigned int, vector unsigned int);
4961 vector float vec_add (vector float, vector float);
4963 vector unsigned int vec_addc (vector unsigned int, vector unsigned int);
4965 vector unsigned char vec_adds (vector signed char,
4966 vector unsigned char);
4967 vector unsigned char vec_adds (vector unsigned char,
4968 vector signed char);
4969 vector unsigned char vec_adds (vector unsigned char,
4970 vector unsigned char);
4971 vector signed char vec_adds (vector signed char, vector signed char);
4972 vector unsigned short vec_adds (vector signed short,
4973 vector unsigned short);
4974 vector unsigned short vec_adds (vector unsigned short,
4975 vector signed short);
4976 vector unsigned short vec_adds (vector unsigned short,
4977 vector unsigned short);
4978 vector signed short vec_adds (vector signed short, vector signed short);
4980 vector unsigned int vec_adds (vector signed int, vector unsigned int);
4981 vector unsigned int vec_adds (vector unsigned int, vector signed int);
4982 vector unsigned int vec_adds (vector unsigned int, vector unsigned int);
4984 vector signed int vec_adds (vector signed int, vector signed int);
4986 vector float vec_and (vector float, vector float);
4987 vector float vec_and (vector float, vector signed int);
4988 vector float vec_and (vector signed int, vector float);
4989 vector signed int vec_and (vector signed int, vector signed int);
4990 vector unsigned int vec_and (vector signed int, vector unsigned int);
4991 vector unsigned int vec_and (vector unsigned int, vector signed int);
4992 vector unsigned int vec_and (vector unsigned int, vector unsigned int);
4993 vector signed short vec_and (vector signed short, vector signed short);
4994 vector unsigned short vec_and (vector signed short,
4995 vector unsigned short);
4996 vector unsigned short vec_and (vector unsigned short,
4997 vector signed short);
4998 vector unsigned short vec_and (vector unsigned short,
4999 vector unsigned short);
5000 vector signed char vec_and (vector signed char, vector signed char);
5001 vector unsigned char vec_and (vector signed char, vector unsigned char);
5003 vector unsigned char vec_and (vector unsigned char, vector signed char);
5005 vector unsigned char vec_and (vector unsigned char,
5006 vector unsigned char);
5008 vector float vec_andc (vector float, vector float);
5009 vector float vec_andc (vector float, vector signed int);
5010 vector float vec_andc (vector signed int, vector float);
5011 vector signed int vec_andc (vector signed int, vector signed int);
5012 vector unsigned int vec_andc (vector signed int, vector unsigned int);
5013 vector unsigned int vec_andc (vector unsigned int, vector signed int);
5014 vector unsigned int vec_andc (vector unsigned int, vector unsigned int);
5016 vector signed short vec_andc (vector signed short, vector signed short);
5018 vector unsigned short vec_andc (vector signed short,
5019 vector unsigned short);
5020 vector unsigned short vec_andc (vector unsigned short,
5021 vector signed short);
5022 vector unsigned short vec_andc (vector unsigned short,
5023 vector unsigned short);
5024 vector signed char vec_andc (vector signed char, vector signed char);
5025 vector unsigned char vec_andc (vector signed char,
5026 vector unsigned char);
5027 vector unsigned char vec_andc (vector unsigned char,
5028 vector signed char);
5029 vector unsigned char vec_andc (vector unsigned char,
5030 vector unsigned char);
5032 vector unsigned char vec_avg (vector unsigned char,
5033 vector unsigned char);
5034 vector signed char vec_avg (vector signed char, vector signed char);
5035 vector unsigned short vec_avg (vector unsigned short,
5036 vector unsigned short);
5037 vector signed short vec_avg (vector signed short, vector signed short);
5038 vector unsigned int vec_avg (vector unsigned int, vector unsigned int);
5039 vector signed int vec_avg (vector signed int, vector signed int);
5041 vector float vec_ceil (vector float);
5043 vector signed int vec_cmpb (vector float, vector float);
5045 vector signed char vec_cmpeq (vector signed char, vector signed char);
5046 vector signed char vec_cmpeq (vector unsigned char,
5047 vector unsigned char);
5048 vector signed short vec_cmpeq (vector signed short,
5049 vector signed short);
5050 vector signed short vec_cmpeq (vector unsigned short,
5051 vector unsigned short);
5052 vector signed int vec_cmpeq (vector signed int, vector signed int);
5053 vector signed int vec_cmpeq (vector unsigned int, vector unsigned int);
5054 vector signed int vec_cmpeq (vector float, vector float);
5056 vector signed int vec_cmpge (vector float, vector float);
5058 vector signed char vec_cmpgt (vector unsigned char,
5059 vector unsigned char);
5060 vector signed char vec_cmpgt (vector signed char, vector signed char);
5061 vector signed short vec_cmpgt (vector unsigned short,
5062 vector unsigned short);
5063 vector signed short vec_cmpgt (vector signed short,
5064 vector signed short);
5065 vector signed int vec_cmpgt (vector unsigned int, vector unsigned int);
5066 vector signed int vec_cmpgt (vector signed int, vector signed int);
5067 vector signed int vec_cmpgt (vector float, vector float);
5069 vector signed int vec_cmple (vector float, vector float);
5071 vector signed char vec_cmplt (vector unsigned char,
5072 vector unsigned char);
5073 vector signed char vec_cmplt (vector signed char, vector signed char);
5074 vector signed short vec_cmplt (vector unsigned short,
5075 vector unsigned short);
5076 vector signed short vec_cmplt (vector signed short,
5077 vector signed short);
5078 vector signed int vec_cmplt (vector unsigned int, vector unsigned int);
5079 vector signed int vec_cmplt (vector signed int, vector signed int);
5080 vector signed int vec_cmplt (vector float, vector float);
5082 vector float vec_ctf (vector unsigned int, const char);
5083 vector float vec_ctf (vector signed int, const char);
5085 vector signed int vec_cts (vector float, const char);
5087 vector unsigned int vec_ctu (vector float, const char);
5089 void vec_dss (const char);
5091 void vec_dssall (void);
5093 void vec_dst (void *, int, const char);
5095 void vec_dstst (void *, int, const char);
5097 void vec_dststt (void *, int, const char);
5099 void vec_dstt (void *, int, const char);
5101 vector float vec_expte (vector float, vector float);
5103 vector float vec_floor (vector float, vector float);
5105 vector float vec_ld (int, vector float *);
5106 vector float vec_ld (int, float *):
5107 vector signed int vec_ld (int, int *);
5108 vector signed int vec_ld (int, vector signed int *);
5109 vector unsigned int vec_ld (int, vector unsigned int *);
5110 vector unsigned int vec_ld (int, unsigned int *);
5111 vector signed short vec_ld (int, short *, vector signed short *);
5112 vector unsigned short vec_ld (int, unsigned short *,
5113 vector unsigned short *);
5114 vector signed char vec_ld (int, signed char *);
5115 vector signed char vec_ld (int, vector signed char *);
5116 vector unsigned char vec_ld (int, unsigned char *);
5117 vector unsigned char vec_ld (int, vector unsigned char *);
5119 vector signed char vec_lde (int, signed char *);
5120 vector unsigned char vec_lde (int, unsigned char *);
5121 vector signed short vec_lde (int, short *);
5122 vector unsigned short vec_lde (int, unsigned short *);
5123 vector float vec_lde (int, float *);
5124 vector signed int vec_lde (int, int *);
5125 vector unsigned int vec_lde (int, unsigned int *);
5127 void float vec_ldl (int, float *);
5128 void float vec_ldl (int, vector float *);
5129 void signed int vec_ldl (int, vector signed int *);
5130 void signed int vec_ldl (int, int *);
5131 void unsigned int vec_ldl (int, unsigned int *);
5132 void unsigned int vec_ldl (int, vector unsigned int *);
5133 void signed short vec_ldl (int, vector signed short *);
5134 void signed short vec_ldl (int, short *);
5135 void unsigned short vec_ldl (int, vector unsigned short *);
5136 void unsigned short vec_ldl (int, unsigned short *);
5137 void signed char vec_ldl (int, vector signed char *);
5138 void signed char vec_ldl (int, signed char *);
5139 void unsigned char vec_ldl (int, vector unsigned char *);
5140 void unsigned char vec_ldl (int, unsigned char *);
5142 vector float vec_loge (vector float);
5144 vector unsigned char vec_lvsl (int, void *, int *);
5146 vector unsigned char vec_lvsr (int, void *, int *);
5148 vector float vec_madd (vector float, vector float, vector float);
5150 vector signed short vec_madds (vector signed short, vector signed short,
5151 vector signed short);
5153 vector unsigned char vec_max (vector signed char, vector unsigned char);
5155 vector unsigned char vec_max (vector unsigned char, vector signed char);
5157 vector unsigned char vec_max (vector unsigned char,
5158 vector unsigned char);
5159 vector signed char vec_max (vector signed char, vector signed char);
5160 vector unsigned short vec_max (vector signed short,
5161 vector unsigned short);
5162 vector unsigned short vec_max (vector unsigned short,
5163 vector signed short);
5164 vector unsigned short vec_max (vector unsigned short,
5165 vector unsigned short);
5166 vector signed short vec_max (vector signed short, vector signed short);
5167 vector unsigned int vec_max (vector signed int, vector unsigned int);
5168 vector unsigned int vec_max (vector unsigned int, vector signed int);
5169 vector unsigned int vec_max (vector unsigned int, vector unsigned int);
5170 vector signed int vec_max (vector signed int, vector signed int);
5171 vector float vec_max (vector float, vector float);
5173 vector signed char vec_mergeh (vector signed char, vector signed char);
5174 vector unsigned char vec_mergeh (vector unsigned char,
5175 vector unsigned char);
5176 vector signed short vec_mergeh (vector signed short,
5177 vector signed short);
5178 vector unsigned short vec_mergeh (vector unsigned short,
5179 vector unsigned short);
5180 vector float vec_mergeh (vector float, vector float);
5181 vector signed int vec_mergeh (vector signed int, vector signed int);
5182 vector unsigned int vec_mergeh (vector unsigned int,
5183 vector unsigned int);
5185 vector signed char vec_mergel (vector signed char, vector signed char);
5186 vector unsigned char vec_mergel (vector unsigned char,
5187 vector unsigned char);
5188 vector signed short vec_mergel (vector signed short,
5189 vector signed short);
5190 vector unsigned short vec_mergel (vector unsigned short,
5191 vector unsigned short);
5192 vector float vec_mergel (vector float, vector float);
5193 vector signed int vec_mergel (vector signed int, vector signed int);
5194 vector unsigned int vec_mergel (vector unsigned int,
5195 vector unsigned int);
5197 vector unsigned short vec_mfvscr (void);
5199 vector unsigned char vec_min (vector signed char, vector unsigned char);
5201 vector unsigned char vec_min (vector unsigned char, vector signed char);
5203 vector unsigned char vec_min (vector unsigned char,
5204 vector unsigned char);
5205 vector signed char vec_min (vector signed char, vector signed char);
5206 vector unsigned short vec_min (vector signed short,
5207 vector unsigned short);
5208 vector unsigned short vec_min (vector unsigned short,
5209 vector signed short);
5210 vector unsigned short vec_min (vector unsigned short,
5211 vector unsigned short);
5212 vector signed short vec_min (vector signed short, vector signed short);
5213 vector unsigned int vec_min (vector signed int, vector unsigned int);
5214 vector unsigned int vec_min (vector unsigned int, vector signed int);
5215 vector unsigned int vec_min (vector unsigned int, vector unsigned int);
5216 vector signed int vec_min (vector signed int, vector signed int);
5217 vector float vec_min (vector float, vector float);
5219 vector signed short vec_mladd (vector signed short, vector signed short,
5220 vector signed short);
5221 vector signed short vec_mladd (vector signed short,
5222 vector unsigned short,
5223 vector unsigned short);
5224 vector signed short vec_mladd (vector unsigned short,
5225 vector signed short,
5226 vector signed short);
5227 vector unsigned short vec_mladd (vector unsigned short,
5228 vector unsigned short,
5229 vector unsigned short);
5231 vector signed short vec_mradds (vector signed short,
5232 vector signed short,
5233 vector signed short);
5235 vector unsigned int vec_msum (vector unsigned char,
5236 vector unsigned char,
5237 vector unsigned int);
5238 vector signed int vec_msum (vector signed char, vector unsigned char,
5240 vector unsigned int vec_msum (vector unsigned short,
5241 vector unsigned short,
5242 vector unsigned int);
5243 vector signed int vec_msum (vector signed short, vector signed short,
5246 vector unsigned int vec_msums (vector unsigned short,
5247 vector unsigned short,
5248 vector unsigned int);
5249 vector signed int vec_msums (vector signed short, vector signed short,
5252 void vec_mtvscr (vector signed int);
5253 void vec_mtvscr (vector unsigned int);
5254 void vec_mtvscr (vector signed short);
5255 void vec_mtvscr (vector unsigned short);
5256 void vec_mtvscr (vector signed char);
5257 void vec_mtvscr (vector unsigned char);
5259 vector unsigned short vec_mule (vector unsigned char,
5260 vector unsigned char);
5261 vector signed short vec_mule (vector signed char, vector signed char);
5262 vector unsigned int vec_mule (vector unsigned short,
5263 vector unsigned short);
5264 vector signed int vec_mule (vector signed short, vector signed short);
5266 vector unsigned short vec_mulo (vector unsigned char,
5267 vector unsigned char);
5268 vector signed short vec_mulo (vector signed char, vector signed char);
5269 vector unsigned int vec_mulo (vector unsigned short,
5270 vector unsigned short);
5271 vector signed int vec_mulo (vector signed short, vector signed short);
5273 vector float vec_nmsub (vector float, vector float, vector float);
5275 vector float vec_nor (vector float, vector float);
5276 vector signed int vec_nor (vector signed int, vector signed int);
5277 vector unsigned int vec_nor (vector unsigned int, vector unsigned int);
5278 vector signed short vec_nor (vector signed short, vector signed short);
5279 vector unsigned short vec_nor (vector unsigned short,
5280 vector unsigned short);
5281 vector signed char vec_nor (vector signed char, vector signed char);
5282 vector unsigned char vec_nor (vector unsigned char,
5283 vector unsigned char);
5285 vector float vec_or (vector float, vector float);
5286 vector float vec_or (vector float, vector signed int);
5287 vector float vec_or (vector signed int, vector float);
5288 vector signed int vec_or (vector signed int, vector signed int);
5289 vector unsigned int vec_or (vector signed int, vector unsigned int);
5290 vector unsigned int vec_or (vector unsigned int, vector signed int);
5291 vector unsigned int vec_or (vector unsigned int, vector unsigned int);
5292 vector signed short vec_or (vector signed short, vector signed short);
5293 vector unsigned short vec_or (vector signed short,
5294 vector unsigned short);
5295 vector unsigned short vec_or (vector unsigned short,
5296 vector signed short);
5297 vector unsigned short vec_or (vector unsigned short,
5298 vector unsigned short);
5299 vector signed char vec_or (vector signed char, vector signed char);
5300 vector unsigned char vec_or (vector signed char, vector unsigned char);
5301 vector unsigned char vec_or (vector unsigned char, vector signed char);
5302 vector unsigned char vec_or (vector unsigned char,
5303 vector unsigned char);
5305 vector signed char vec_pack (vector signed short, vector signed short);
5306 vector unsigned char vec_pack (vector unsigned short,
5307 vector unsigned short);
5308 vector signed short vec_pack (vector signed int, vector signed int);
5309 vector unsigned short vec_pack (vector unsigned int,
5310 vector unsigned int);
5312 vector signed short vec_packpx (vector unsigned int,
5313 vector unsigned int);
5315 vector unsigned char vec_packs (vector unsigned short,
5316 vector unsigned short);
5317 vector signed char vec_packs (vector signed short, vector signed short);
5319 vector unsigned short vec_packs (vector unsigned int,
5320 vector unsigned int);
5321 vector signed short vec_packs (vector signed int, vector signed int);
5323 vector unsigned char vec_packsu (vector unsigned short,
5324 vector unsigned short);
5325 vector unsigned char vec_packsu (vector signed short,
5326 vector signed short);
5327 vector unsigned short vec_packsu (vector unsigned int,
5328 vector unsigned int);
5329 vector unsigned short vec_packsu (vector signed int, vector signed int);
5331 vector float vec_perm (vector float, vector float,
5332 vector unsigned char);
5333 vector signed int vec_perm (vector signed int, vector signed int,
5334 vector unsigned char);
5335 vector unsigned int vec_perm (vector unsigned int, vector unsigned int,
5336 vector unsigned char);
5337 vector signed short vec_perm (vector signed short, vector signed short,
5338 vector unsigned char);
5339 vector unsigned short vec_perm (vector unsigned short,
5340 vector unsigned short,
5341 vector unsigned char);
5342 vector signed char vec_perm (vector signed char, vector signed char,
5343 vector unsigned char);
5344 vector unsigned char vec_perm (vector unsigned char,
5345 vector unsigned char,
5346 vector unsigned char);
5348 vector float vec_re (vector float);
5350 vector signed char vec_rl (vector signed char, vector unsigned char);
5351 vector unsigned char vec_rl (vector unsigned char,
5352 vector unsigned char);
5353 vector signed short vec_rl (vector signed short, vector unsigned short);
5355 vector unsigned short vec_rl (vector unsigned short,
5356 vector unsigned short);
5357 vector signed int vec_rl (vector signed int, vector unsigned int);
5358 vector unsigned int vec_rl (vector unsigned int, vector unsigned int);
5360 vector float vec_round (vector float);
5362 vector float vec_rsqrte (vector float);
5364 vector float vec_sel (vector float, vector float, vector signed int);
5365 vector float vec_sel (vector float, vector float, vector unsigned int);
5366 vector signed int vec_sel (vector signed int, vector signed int,
5368 vector signed int vec_sel (vector signed int, vector signed int,
5369 vector unsigned int);
5370 vector unsigned int vec_sel (vector unsigned int, vector unsigned int,
5372 vector unsigned int vec_sel (vector unsigned int, vector unsigned int,
5373 vector unsigned int);
5374 vector signed short vec_sel (vector signed short, vector signed short,
5375 vector signed short);
5376 vector signed short vec_sel (vector signed short, vector signed short,
5377 vector unsigned short);
5378 vector unsigned short vec_sel (vector unsigned short,
5379 vector unsigned short,
5380 vector signed short);
5381 vector unsigned short vec_sel (vector unsigned short,
5382 vector unsigned short,
5383 vector unsigned short);
5384 vector signed char vec_sel (vector signed char, vector signed char,
5385 vector signed char);
5386 vector signed char vec_sel (vector signed char, vector signed char,
5387 vector unsigned char);
5388 vector unsigned char vec_sel (vector unsigned char,
5389 vector unsigned char,
5390 vector signed char);
5391 vector unsigned char vec_sel (vector unsigned char,
5392 vector unsigned char,
5393 vector unsigned char);
5395 vector signed char vec_sl (vector signed char, vector unsigned char);
5396 vector unsigned char vec_sl (vector unsigned char,
5397 vector unsigned char);
5398 vector signed short vec_sl (vector signed short, vector unsigned short);
5400 vector unsigned short vec_sl (vector unsigned short,
5401 vector unsigned short);
5402 vector signed int vec_sl (vector signed int, vector unsigned int);
5403 vector unsigned int vec_sl (vector unsigned int, vector unsigned int);
5405 vector float vec_sld (vector float, vector float, const char);
5406 vector signed int vec_sld (vector signed int, vector signed int,
5408 vector unsigned int vec_sld (vector unsigned int, vector unsigned int,
5410 vector signed short vec_sld (vector signed short, vector signed short,
5412 vector unsigned short vec_sld (vector unsigned short,
5413 vector unsigned short, const char);
5414 vector signed char vec_sld (vector signed char, vector signed char,
5416 vector unsigned char vec_sld (vector unsigned char,
5417 vector unsigned char,
5420 vector signed int vec_sll (vector signed int, vector unsigned int);
5421 vector signed int vec_sll (vector signed int, vector unsigned short);
5422 vector signed int vec_sll (vector signed int, vector unsigned char);
5423 vector unsigned int vec_sll (vector unsigned int, vector unsigned int);
5424 vector unsigned int vec_sll (vector unsigned int,
5425 vector unsigned short);
5426 vector unsigned int vec_sll (vector unsigned int, vector unsigned char);
5428 vector signed short vec_sll (vector signed short, vector unsigned int);
5429 vector signed short vec_sll (vector signed short,
5430 vector unsigned short);
5431 vector signed short vec_sll (vector signed short, vector unsigned char);
5433 vector unsigned short vec_sll (vector unsigned short,
5434 vector unsigned int);
5435 vector unsigned short vec_sll (vector unsigned short,
5436 vector unsigned short);
5437 vector unsigned short vec_sll (vector unsigned short,
5438 vector unsigned char);
5439 vector signed char vec_sll (vector signed char, vector unsigned int);
5440 vector signed char vec_sll (vector signed char, vector unsigned short);
5441 vector signed char vec_sll (vector signed char, vector unsigned char);
5442 vector unsigned char vec_sll (vector unsigned char,
5443 vector unsigned int);
5444 vector unsigned char vec_sll (vector unsigned char,
5445 vector unsigned short);
5446 vector unsigned char vec_sll (vector unsigned char,
5447 vector unsigned char);
5449 vector float vec_slo (vector float, vector signed char);
5450 vector float vec_slo (vector float, vector unsigned char);
5451 vector signed int vec_slo (vector signed int, vector signed char);
5452 vector signed int vec_slo (vector signed int, vector unsigned char);
5453 vector unsigned int vec_slo (vector unsigned int, vector signed char);
5454 vector unsigned int vec_slo (vector unsigned int, vector unsigned char);
5456 vector signed short vec_slo (vector signed short, vector signed char);
5457 vector signed short vec_slo (vector signed short, vector unsigned char);
5459 vector unsigned short vec_slo (vector unsigned short,
5460 vector signed char);
5461 vector unsigned short vec_slo (vector unsigned short,
5462 vector unsigned char);
5463 vector signed char vec_slo (vector signed char, vector signed char);
5464 vector signed char vec_slo (vector signed char, vector unsigned char);
5465 vector unsigned char vec_slo (vector unsigned char, vector signed char);
5467 vector unsigned char vec_slo (vector unsigned char,
5468 vector unsigned char);
5470 vector signed char vec_splat (vector signed char, const char);
5471 vector unsigned char vec_splat (vector unsigned char, const char);
5472 vector signed short vec_splat (vector signed short, const char);
5473 vector unsigned short vec_splat (vector unsigned short, const char);
5474 vector float vec_splat (vector float, const char);
5475 vector signed int vec_splat (vector signed int, const char);
5476 vector unsigned int vec_splat (vector unsigned int, const char);
5478 vector signed char vec_splat_s8 (const char);
5480 vector signed short vec_splat_s16 (const char);
5482 vector signed int vec_splat_s32 (const char);
5484 vector unsigned char vec_splat_u8 (const char);
5486 vector unsigned short vec_splat_u16 (const char);
5488 vector unsigned int vec_splat_u32 (const char);
5490 vector signed char vec_sr (vector signed char, vector unsigned char);
5491 vector unsigned char vec_sr (vector unsigned char,
5492 vector unsigned char);
5493 vector signed short vec_sr (vector signed short, vector unsigned short);
5495 vector unsigned short vec_sr (vector unsigned short,
5496 vector unsigned short);
5497 vector signed int vec_sr (vector signed int, vector unsigned int);
5498 vector unsigned int vec_sr (vector unsigned int, vector unsigned int);
5500 vector signed char vec_sra (vector signed char, vector unsigned char);
5501 vector unsigned char vec_sra (vector unsigned char,
5502 vector unsigned char);
5503 vector signed short vec_sra (vector signed short,
5504 vector unsigned short);
5505 vector unsigned short vec_sra (vector unsigned short,
5506 vector unsigned short);
5507 vector signed int vec_sra (vector signed int, vector unsigned int);
5508 vector unsigned int vec_sra (vector unsigned int, vector unsigned int);
5510 vector signed int vec_srl (vector signed int, vector unsigned int);
5511 vector signed int vec_srl (vector signed int, vector unsigned short);
5512 vector signed int vec_srl (vector signed int, vector unsigned char);
5513 vector unsigned int vec_srl (vector unsigned int, vector unsigned int);
5514 vector unsigned int vec_srl (vector unsigned int,
5515 vector unsigned short);
5516 vector unsigned int vec_srl (vector unsigned int, vector unsigned char);
5518 vector signed short vec_srl (vector signed short, vector unsigned int);
5519 vector signed short vec_srl (vector signed short,
5520 vector unsigned short);
5521 vector signed short vec_srl (vector signed short, vector unsigned char);
5523 vector unsigned short vec_srl (vector unsigned short,
5524 vector unsigned int);
5525 vector unsigned short vec_srl (vector unsigned short,
5526 vector unsigned short);
5527 vector unsigned short vec_srl (vector unsigned short,
5528 vector unsigned char);
5529 vector signed char vec_srl (vector signed char, vector unsigned int);
5530 vector signed char vec_srl (vector signed char, vector unsigned short);
5531 vector signed char vec_srl (vector signed char, vector unsigned char);
5532 vector unsigned char vec_srl (vector unsigned char,
5533 vector unsigned int);
5534 vector unsigned char vec_srl (vector unsigned char,
5535 vector unsigned short);
5536 vector unsigned char vec_srl (vector unsigned char,
5537 vector unsigned char);
5539 vector float vec_sro (vector float, vector signed char);
5540 vector float vec_sro (vector float, vector unsigned char);
5541 vector signed int vec_sro (vector signed int, vector signed char);
5542 vector signed int vec_sro (vector signed int, vector unsigned char);
5543 vector unsigned int vec_sro (vector unsigned int, vector signed char);
5544 vector unsigned int vec_sro (vector unsigned int, vector unsigned char);
5546 vector signed short vec_sro (vector signed short, vector signed char);
5547 vector signed short vec_sro (vector signed short, vector unsigned char);
5549 vector unsigned short vec_sro (vector unsigned short,
5550 vector signed char);
5551 vector unsigned short vec_sro (vector unsigned short,
5552 vector unsigned char);
5553 vector signed char vec_sro (vector signed char, vector signed char);
5554 vector signed char vec_sro (vector signed char, vector unsigned char);
5555 vector unsigned char vec_sro (vector unsigned char, vector signed char);
5557 vector unsigned char vec_sro (vector unsigned char,
5558 vector unsigned char);
5560 void vec_st (vector float, int, float *);
5561 void vec_st (vector float, int, vector float *);
5562 void vec_st (vector signed int, int, int *);
5563 void vec_st (vector signed int, int, unsigned int *);
5564 void vec_st (vector unsigned int, int, unsigned int *);
5565 void vec_st (vector unsigned int, int, vector unsigned int *);
5566 void vec_st (vector signed short, int, short *);
5567 void vec_st (vector signed short, int, vector unsigned short *);
5568 void vec_st (vector signed short, int, vector signed short *);
5569 void vec_st (vector unsigned short, int, unsigned short *);
5570 void vec_st (vector unsigned short, int, vector unsigned short *);
5571 void vec_st (vector signed char, int, signed char *);
5572 void vec_st (vector signed char, int, unsigned char *);
5573 void vec_st (vector signed char, int, vector signed char *);
5574 void vec_st (vector unsigned char, int, unsigned char *);
5575 void vec_st (vector unsigned char, int, vector unsigned char *);
5577 void vec_ste (vector signed char, int, unsigned char *);
5578 void vec_ste (vector signed char, int, signed char *);
5579 void vec_ste (vector unsigned char, int, unsigned char *);
5580 void vec_ste (vector signed short, int, short *);
5581 void vec_ste (vector signed short, int, unsigned short *);
5582 void vec_ste (vector unsigned short, int, void *);
5583 void vec_ste (vector signed int, int, unsigned int *);
5584 void vec_ste (vector signed int, int, int *);
5585 void vec_ste (vector unsigned int, int, unsigned int *);
5586 void vec_ste (vector float, int, float *);
5588 void vec_stl (vector float, int, vector float *);
5589 void vec_stl (vector float, int, float *);
5590 void vec_stl (vector signed int, int, vector signed int *);
5591 void vec_stl (vector signed int, int, int *);
5592 void vec_stl (vector signed int, int, unsigned int *);
5593 void vec_stl (vector unsigned int, int, vector unsigned int *);
5594 void vec_stl (vector unsigned int, int, unsigned int *);
5595 void vec_stl (vector signed short, int, short *);
5596 void vec_stl (vector signed short, int, unsigned short *);
5597 void vec_stl (vector signed short, int, vector signed short *);
5598 void vec_stl (vector unsigned short, int, unsigned short *);
5599 void vec_stl (vector unsigned short, int, vector signed short *);
5600 void vec_stl (vector signed char, int, signed char *);
5601 void vec_stl (vector signed char, int, unsigned char *);
5602 void vec_stl (vector signed char, int, vector signed char *);
5603 void vec_stl (vector unsigned char, int, unsigned char *);
5604 void vec_stl (vector unsigned char, int, vector unsigned char *);
5606 vector signed char vec_sub (vector signed char, vector signed char);
5607 vector unsigned char vec_sub (vector signed char, vector unsigned char);
5609 vector unsigned char vec_sub (vector unsigned char, vector signed char);
5611 vector unsigned char vec_sub (vector unsigned char,
5612 vector unsigned char);
5613 vector signed short vec_sub (vector signed short, vector signed short);
5614 vector unsigned short vec_sub (vector signed short,
5615 vector unsigned short);
5616 vector unsigned short vec_sub (vector unsigned short,
5617 vector signed short);
5618 vector unsigned short vec_sub (vector unsigned short,
5619 vector unsigned short);
5620 vector signed int vec_sub (vector signed int, vector signed int);
5621 vector unsigned int vec_sub (vector signed int, vector unsigned int);
5622 vector unsigned int vec_sub (vector unsigned int, vector signed int);
5623 vector unsigned int vec_sub (vector unsigned int, vector unsigned int);
5624 vector float vec_sub (vector float, vector float);
5626 vector unsigned int vec_subc (vector unsigned int, vector unsigned int);
5628 vector unsigned char vec_subs (vector signed char,
5629 vector unsigned char);
5630 vector unsigned char vec_subs (vector unsigned char,
5631 vector signed char);
5632 vector unsigned char vec_subs (vector unsigned char,
5633 vector unsigned char);
5634 vector signed char vec_subs (vector signed char, vector signed char);
5635 vector unsigned short vec_subs (vector signed short,
5636 vector unsigned short);
5637 vector unsigned short vec_subs (vector unsigned short,
5638 vector signed short);
5639 vector unsigned short vec_subs (vector unsigned short,
5640 vector unsigned short);
5641 vector signed short vec_subs (vector signed short, vector signed short);
5643 vector unsigned int vec_subs (vector signed int, vector unsigned int);
5644 vector unsigned int vec_subs (vector unsigned int, vector signed int);
5645 vector unsigned int vec_subs (vector unsigned int, vector unsigned int);
5647 vector signed int vec_subs (vector signed int, vector signed int);
5649 vector unsigned int vec_sum4s (vector unsigned char,
5650 vector unsigned int);
5651 vector signed int vec_sum4s (vector signed char, vector signed int);
5652 vector signed int vec_sum4s (vector signed short, vector signed int);
5654 vector signed int vec_sum2s (vector signed int, vector signed int);
5656 vector signed int vec_sums (vector signed int, vector signed int);
5658 vector float vec_trunc (vector float);
5660 vector signed short vec_unpackh (vector signed char);
5661 vector unsigned int vec_unpackh (vector signed short);
5662 vector signed int vec_unpackh (vector signed short);
5664 vector signed short vec_unpackl (vector signed char);
5665 vector unsigned int vec_unpackl (vector signed short);
5666 vector signed int vec_unpackl (vector signed short);
5668 vector float vec_xor (vector float, vector float);
5669 vector float vec_xor (vector float, vector signed int);
5670 vector float vec_xor (vector signed int, vector float);
5671 vector signed int vec_xor (vector signed int, vector signed int);
5672 vector unsigned int vec_xor (vector signed int, vector unsigned int);
5673 vector unsigned int vec_xor (vector unsigned int, vector signed int);
5674 vector unsigned int vec_xor (vector unsigned int, vector unsigned int);
5675 vector signed short vec_xor (vector signed short, vector signed short);
5676 vector unsigned short vec_xor (vector signed short,
5677 vector unsigned short);
5678 vector unsigned short vec_xor (vector unsigned short,
5679 vector signed short);
5680 vector unsigned short vec_xor (vector unsigned short,
5681 vector unsigned short);
5682 vector signed char vec_xor (vector signed char, vector signed char);
5683 vector unsigned char vec_xor (vector signed char, vector unsigned char);
5685 vector unsigned char vec_xor (vector unsigned char, vector signed char);
5687 vector unsigned char vec_xor (vector unsigned char,
5688 vector unsigned char);
5690 vector signed int vec_all_eq (vector signed char, vector unsigned char);
5692 vector signed int vec_all_eq (vector signed char, vector signed char);
5693 vector signed int vec_all_eq (vector unsigned char, vector signed char);
5695 vector signed int vec_all_eq (vector unsigned char,
5696 vector unsigned char);
5697 vector signed int vec_all_eq (vector signed short,
5698 vector unsigned short);
5699 vector signed int vec_all_eq (vector signed short, vector signed short);
5701 vector signed int vec_all_eq (vector unsigned short,
5702 vector signed short);
5703 vector signed int vec_all_eq (vector unsigned short,
5704 vector unsigned short);
5705 vector signed int vec_all_eq (vector signed int, vector unsigned int);
5706 vector signed int vec_all_eq (vector signed int, vector signed int);
5707 vector signed int vec_all_eq (vector unsigned int, vector signed int);
5708 vector signed int vec_all_eq (vector unsigned int, vector unsigned int);
5710 vector signed int vec_all_eq (vector float, vector float);
5712 vector signed int vec_all_ge (vector signed char, vector unsigned char);
5714 vector signed int vec_all_ge (vector unsigned char, vector signed char);
5716 vector signed int vec_all_ge (vector unsigned char,
5717 vector unsigned char);
5718 vector signed int vec_all_ge (vector signed char, vector signed char);
5719 vector signed int vec_all_ge (vector signed short,
5720 vector unsigned short);
5721 vector signed int vec_all_ge (vector unsigned short,
5722 vector signed short);
5723 vector signed int vec_all_ge (vector unsigned short,
5724 vector unsigned short);
5725 vector signed int vec_all_ge (vector signed short, vector signed short);
5727 vector signed int vec_all_ge (vector signed int, vector unsigned int);
5728 vector signed int vec_all_ge (vector unsigned int, vector signed int);
5729 vector signed int vec_all_ge (vector unsigned int, vector unsigned int);
5731 vector signed int vec_all_ge (vector signed int, vector signed int);
5732 vector signed int vec_all_ge (vector float, vector float);
5734 vector signed int vec_all_gt (vector signed char, vector unsigned char);
5736 vector signed int vec_all_gt (vector unsigned char, vector signed char);
5738 vector signed int vec_all_gt (vector unsigned char,
5739 vector unsigned char);
5740 vector signed int vec_all_gt (vector signed char, vector signed char);
5741 vector signed int vec_all_gt (vector signed short,
5742 vector unsigned short);
5743 vector signed int vec_all_gt (vector unsigned short,
5744 vector signed short);
5745 vector signed int vec_all_gt (vector unsigned short,
5746 vector unsigned short);
5747 vector signed int vec_all_gt (vector signed short, vector signed short);
5749 vector signed int vec_all_gt (vector signed int, vector unsigned int);
5750 vector signed int vec_all_gt (vector unsigned int, vector signed int);
5751 vector signed int vec_all_gt (vector unsigned int, vector unsigned int);
5753 vector signed int vec_all_gt (vector signed int, vector signed int);
5754 vector signed int vec_all_gt (vector float, vector float);
5756 vector signed int vec_all_in (vector float, vector float);
5758 vector signed int vec_all_le (vector signed char, vector unsigned char);
5760 vector signed int vec_all_le (vector unsigned char, vector signed char);
5762 vector signed int vec_all_le (vector unsigned char,
5763 vector unsigned char);
5764 vector signed int vec_all_le (vector signed char, vector signed char);
5765 vector signed int vec_all_le (vector signed short,
5766 vector unsigned short);
5767 vector signed int vec_all_le (vector unsigned short,
5768 vector signed short);
5769 vector signed int vec_all_le (vector unsigned short,
5770 vector unsigned short);
5771 vector signed int vec_all_le (vector signed short, vector signed short);
5773 vector signed int vec_all_le (vector signed int, vector unsigned int);
5774 vector signed int vec_all_le (vector unsigned int, vector signed int);
5775 vector signed int vec_all_le (vector unsigned int, vector unsigned int);
5777 vector signed int vec_all_le (vector signed int, vector signed int);
5778 vector signed int vec_all_le (vector float, vector float);
5780 vector signed int vec_all_lt (vector signed char, vector unsigned char);
5782 vector signed int vec_all_lt (vector unsigned char, vector signed char);
5784 vector signed int vec_all_lt (vector unsigned char,
5785 vector unsigned char);
5786 vector signed int vec_all_lt (vector signed char, vector signed char);
5787 vector signed int vec_all_lt (vector signed short,
5788 vector unsigned short);
5789 vector signed int vec_all_lt (vector unsigned short,
5790 vector signed short);
5791 vector signed int vec_all_lt (vector unsigned short,
5792 vector unsigned short);
5793 vector signed int vec_all_lt (vector signed short, vector signed short);
5795 vector signed int vec_all_lt (vector signed int, vector unsigned int);
5796 vector signed int vec_all_lt (vector unsigned int, vector signed int);
5797 vector signed int vec_all_lt (vector unsigned int, vector unsigned int);
5799 vector signed int vec_all_lt (vector signed int, vector signed int);
5800 vector signed int vec_all_lt (vector float, vector float);
5802 vector signed int vec_all_nan (vector float);
5804 vector signed int vec_all_ne (vector signed char, vector unsigned char);
5806 vector signed int vec_all_ne (vector signed char, vector signed char);
5807 vector signed int vec_all_ne (vector unsigned char, vector signed char);
5809 vector signed int vec_all_ne (vector unsigned char,
5810 vector unsigned char);
5811 vector signed int vec_all_ne (vector signed short,
5812 vector unsigned short);
5813 vector signed int vec_all_ne (vector signed short, vector signed short);
5815 vector signed int vec_all_ne (vector unsigned short,
5816 vector signed short);
5817 vector signed int vec_all_ne (vector unsigned short,
5818 vector unsigned short);
5819 vector signed int vec_all_ne (vector signed int, vector unsigned int);
5820 vector signed int vec_all_ne (vector signed int, vector signed int);
5821 vector signed int vec_all_ne (vector unsigned int, vector signed int);
5822 vector signed int vec_all_ne (vector unsigned int, vector unsigned int);
5824 vector signed int vec_all_ne (vector float, vector float);
5826 vector signed int vec_all_nge (vector float, vector float);
5828 vector signed int vec_all_ngt (vector float, vector float);
5830 vector signed int vec_all_nle (vector float, vector float);
5832 vector signed int vec_all_nlt (vector float, vector float);
5834 vector signed int vec_all_numeric (vector float);
5836 vector signed int vec_any_eq (vector signed char, vector unsigned char);
5838 vector signed int vec_any_eq (vector signed char, vector signed char);
5839 vector signed int vec_any_eq (vector unsigned char, vector signed char);
5841 vector signed int vec_any_eq (vector unsigned char,
5842 vector unsigned char);
5843 vector signed int vec_any_eq (vector signed short,
5844 vector unsigned short);
5845 vector signed int vec_any_eq (vector signed short, vector signed short);
5847 vector signed int vec_any_eq (vector unsigned short,
5848 vector signed short);
5849 vector signed int vec_any_eq (vector unsigned short,
5850 vector unsigned short);
5851 vector signed int vec_any_eq (vector signed int, vector unsigned int);
5852 vector signed int vec_any_eq (vector signed int, vector signed int);
5853 vector signed int vec_any_eq (vector unsigned int, vector signed int);
5854 vector signed int vec_any_eq (vector unsigned int, vector unsigned int);
5856 vector signed int vec_any_eq (vector float, vector float);
5858 vector signed int vec_any_ge (vector signed char, vector unsigned char);
5860 vector signed int vec_any_ge (vector unsigned char, vector signed char);
5862 vector signed int vec_any_ge (vector unsigned char,
5863 vector unsigned char);
5864 vector signed int vec_any_ge (vector signed char, vector signed char);
5865 vector signed int vec_any_ge (vector signed short,
5866 vector unsigned short);
5867 vector signed int vec_any_ge (vector unsigned short,
5868 vector signed short);
5869 vector signed int vec_any_ge (vector unsigned short,
5870 vector unsigned short);
5871 vector signed int vec_any_ge (vector signed short, vector signed short);
5873 vector signed int vec_any_ge (vector signed int, vector unsigned int);
5874 vector signed int vec_any_ge (vector unsigned int, vector signed int);
5875 vector signed int vec_any_ge (vector unsigned int, vector unsigned int);
5877 vector signed int vec_any_ge (vector signed int, vector signed int);
5878 vector signed int vec_any_ge (vector float, vector float);
5880 vector signed int vec_any_gt (vector signed char, vector unsigned char);
5882 vector signed int vec_any_gt (vector unsigned char, vector signed char);
5884 vector signed int vec_any_gt (vector unsigned char,
5885 vector unsigned char);
5886 vector signed int vec_any_gt (vector signed char, vector signed char);
5887 vector signed int vec_any_gt (vector signed short,
5888 vector unsigned short);
5889 vector signed int vec_any_gt (vector unsigned short,
5890 vector signed short);
5891 vector signed int vec_any_gt (vector unsigned short,
5892 vector unsigned short);
5893 vector signed int vec_any_gt (vector signed short, vector signed short);
5895 vector signed int vec_any_gt (vector signed int, vector unsigned int);
5896 vector signed int vec_any_gt (vector unsigned int, vector signed int);
5897 vector signed int vec_any_gt (vector unsigned int, vector unsigned int);
5899 vector signed int vec_any_gt (vector signed int, vector signed int);
5900 vector signed int vec_any_gt (vector float, vector float);
5902 vector signed int vec_any_le (vector signed char, vector unsigned char);
5904 vector signed int vec_any_le (vector unsigned char, vector signed char);
5906 vector signed int vec_any_le (vector unsigned char,
5907 vector unsigned char);
5908 vector signed int vec_any_le (vector signed char, vector signed char);
5909 vector signed int vec_any_le (vector signed short,
5910 vector unsigned short);
5911 vector signed int vec_any_le (vector unsigned short,
5912 vector signed short);
5913 vector signed int vec_any_le (vector unsigned short,
5914 vector unsigned short);
5915 vector signed int vec_any_le (vector signed short, vector signed short);
5917 vector signed int vec_any_le (vector signed int, vector unsigned int);
5918 vector signed int vec_any_le (vector unsigned int, vector signed int);
5919 vector signed int vec_any_le (vector unsigned int, vector unsigned int);
5921 vector signed int vec_any_le (vector signed int, vector signed int);
5922 vector signed int vec_any_le (vector float, vector float);
5924 vector signed int vec_any_lt (vector signed char, vector unsigned char);
5926 vector signed int vec_any_lt (vector unsigned char, vector signed char);
5928 vector signed int vec_any_lt (vector unsigned char,
5929 vector unsigned char);
5930 vector signed int vec_any_lt (vector signed char, vector signed char);
5931 vector signed int vec_any_lt (vector signed short,
5932 vector unsigned short);
5933 vector signed int vec_any_lt (vector unsigned short,
5934 vector signed short);
5935 vector signed int vec_any_lt (vector unsigned short,
5936 vector unsigned short);
5937 vector signed int vec_any_lt (vector signed short, vector signed short);
5939 vector signed int vec_any_lt (vector signed int, vector unsigned int);
5940 vector signed int vec_any_lt (vector unsigned int, vector signed int);
5941 vector signed int vec_any_lt (vector unsigned int, vector unsigned int);
5943 vector signed int vec_any_lt (vector signed int, vector signed int);
5944 vector signed int vec_any_lt (vector float, vector float);
5946 vector signed int vec_any_nan (vector float);
5948 vector signed int vec_any_ne (vector signed char, vector unsigned char);
5950 vector signed int vec_any_ne (vector signed char, vector signed char);
5951 vector signed int vec_any_ne (vector unsigned char, vector signed char);
5953 vector signed int vec_any_ne (vector unsigned char,
5954 vector unsigned char);
5955 vector signed int vec_any_ne (vector signed short,
5956 vector unsigned short);
5957 vector signed int vec_any_ne (vector signed short, vector signed short);
5959 vector signed int vec_any_ne (vector unsigned short,
5960 vector signed short);
5961 vector signed int vec_any_ne (vector unsigned short,
5962 vector unsigned short);
5963 vector signed int vec_any_ne (vector signed int, vector unsigned int);
5964 vector signed int vec_any_ne (vector signed int, vector signed int);
5965 vector signed int vec_any_ne (vector unsigned int, vector signed int);
5966 vector signed int vec_any_ne (vector unsigned int, vector unsigned int);
5968 vector signed int vec_any_ne (vector float, vector float);
5970 vector signed int vec_any_nge (vector float, vector float);
5972 vector signed int vec_any_ngt (vector float, vector float);
5974 vector signed int vec_any_nle (vector float, vector float);
5976 vector signed int vec_any_nlt (vector float, vector float);
5978 vector signed int vec_any_numeric (vector float);
5980 vector signed int vec_any_out (vector float, vector float);
5984 @section Pragmas Accepted by GCC
5988 GCC supports several types of pragmas, primarily in order to compile
5989 code originally written for other compilers. Note that in general
5990 we do not recommend the use of pragmas; @xref{Function Attributes},
5991 for further explanation.
6001 @subsection ARM Pragmas
6003 The ARM target defines pragmas for controlling the default addition of
6004 @code{long_call} and @code{short_call} attributes to functions.
6005 @xref{Function Attributes}, for information about the effects of these
6010 @cindex pragma, long_calls
6011 Set all subsequent functions to have the @code{long_call} attribute.
6014 @cindex pragma, no_long_calls
6015 Set all subsequent functions to have the @code{short_call} attribute.
6017 @item long_calls_off
6018 @cindex pragma, long_calls_off
6019 Do not affect the @code{long_call} or @code{short_call} attributes of
6020 subsequent functions.
6023 @c Describe c4x pragmas here.
6024 @c Describe h8300 pragmas here.
6025 @c Describe i370 pragmas here.
6026 @c Describe i960 pragmas here.
6027 @c Describe sh pragmas here.
6028 @c Describe v850 pragmas here.
6030 @node Darwin Pragmas
6031 @subsection Darwin Pragmas
6033 The following pragmas are available for all architectures running the
6034 Darwin operating system. These are useful for compatibility with other
6038 @item mark @var{tokens}@dots{}
6039 @cindex pragma, mark
6040 This pragma is accepted, but has no effect.
6042 @item options align=@var{alignment}
6043 @cindex pragma, options align
6044 This pragma sets the alignment of fields in structures. The values of
6045 @var{alignment} may be @code{mac68k}, to emulate m68k alignment, or
6046 @code{power}, to emulate PowerPC alignment. Uses of this pragma nest
6047 properly; to restore the previous setting, use @code{reset} for the
6050 @item segment @var{tokens}@dots{}
6051 @cindex pragma, segment
6052 This pragma is accepted, but has no effect.
6054 @item unused (@var{var} [, @var{var}]@dots{})
6055 @cindex pragma, unused
6056 This pragma declares variables to be possibly unused. GCC will not
6057 produce warnings for the listed variables. The effect is similar to
6058 that of the @code{unused} attribute, except that this pragma may appear
6059 anywhere within the variables' scopes.
6062 @node Solaris Pragmas
6063 @subsection Solaris Pragmas
6065 For compatibility with the SunPRO compiler, the following pragma
6069 @item redefine_extname @var{oldname} @var{newname}
6070 @cindex pragma, redefine_extname
6072 This pragma gives the C function @var{oldname} the assembler label
6073 @var{newname}. The pragma must appear before the function declaration.
6074 This pragma is equivalent to the asm labels extension (@pxref{Asm
6075 Labels}). The preprocessor defines @code{__PRAGMA_REDEFINE_EXTNAME}
6076 if the pragma is available.
6080 @subsection Tru64 Pragmas
6082 For compatibility with the Compaq C compiler, the following pragma
6086 @item extern_prefix @var{string}
6087 @cindex pragma, extern_prefix
6089 This pragma renames all subsequent function and variable declarations
6090 such that @var{string} is prepended to the name. This effect may be
6091 terminated by using another @code{extern_prefix} pragma with the
6094 This pragma is similar in intent to to the asm labels extension
6095 (@pxref{Asm Labels}) in that the system programmer wants to change
6096 the assembly-level ABI without changing the source-level API. The
6097 preprocessor defines @code{__EXTERN_PREFIX} if the pragma is available.
6100 @node Unnamed Fields
6101 @section Unnamed struct/union fields within structs/unions.
6105 For compatibility with other compilers, GCC allows you to define
6106 a structure or union that contains, as fields, structures and unions
6107 without names. For example:
6120 In this example, the user would be able to access members of the unnamed
6121 union with code like @samp{foo.b}. Note that only unnamed structs and
6122 unions are allowed, you may not have, for example, an unnamed
6125 You must never create such structures that cause ambiguous field definitions.
6126 For example, this structure:
6137 It is ambiguous which @code{a} is being referred to with @samp{foo.a}.
6138 Such constructs are not supported and must be avoided. In the future,
6139 such constructs may be detected and treated as compilation errors.
6141 @node C++ Extensions
6142 @chapter Extensions to the C++ Language
6143 @cindex extensions, C++ language
6144 @cindex C++ language extensions
6146 The GNU compiler provides these extensions to the C++ language (and you
6147 can also use most of the C language extensions in your C++ programs). If you
6148 want to write code that checks whether these features are available, you can
6149 test for the GNU compiler the same way as for C programs: check for a
6150 predefined macro @code{__GNUC__}. You can also use @code{__GNUG__} to
6151 test specifically for GNU C++ (@pxref{Standard Predefined,,Standard
6152 Predefined Macros,cpp.info,The C Preprocessor}).
6155 * Min and Max:: C++ Minimum and maximum operators.
6156 * Volatiles:: What constitutes an access to a volatile object.
6157 * Restricted Pointers:: C99 restricted pointers and references.
6158 * Vague Linkage:: Where G++ puts inlines, vtables and such.
6159 * C++ Interface:: You can use a single C++ header file for both
6160 declarations and definitions.
6161 * Template Instantiation:: Methods for ensuring that exactly one copy of
6162 each needed template instantiation is emitted.
6163 * Bound member functions:: You can extract a function pointer to the
6164 method denoted by a @samp{->*} or @samp{.*} expression.
6165 * C++ Attributes:: Variable, function, and type attributes for C++ only.
6166 * Java Exceptions:: Tweaking exception handling to work with Java.
6167 * Deprecated Features:: Things might disappear from g++.
6168 * Backwards Compatibility:: Compatibilities with earlier definitions of C++.
6172 @section Minimum and Maximum Operators in C++
6174 It is very convenient to have operators which return the ``minimum'' or the
6175 ``maximum'' of two arguments. In GNU C++ (but not in GNU C),
6178 @item @var{a} <? @var{b}
6180 @cindex minimum operator
6181 is the @dfn{minimum}, returning the smaller of the numeric values
6182 @var{a} and @var{b};
6184 @item @var{a} >? @var{b}
6186 @cindex maximum operator
6187 is the @dfn{maximum}, returning the larger of the numeric values @var{a}
6191 These operations are not primitive in ordinary C++, since you can
6192 use a macro to return the minimum of two things in C++, as in the
6196 #define MIN(X,Y) ((X) < (Y) ? : (X) : (Y))
6200 You might then use @w{@samp{int min = MIN (i, j);}} to set @var{min} to
6201 the minimum value of variables @var{i} and @var{j}.
6203 However, side effects in @code{X} or @code{Y} may cause unintended
6204 behavior. For example, @code{MIN (i++, j++)} will fail, incrementing
6205 the smaller counter twice. A GNU C extension allows you to write safe
6206 macros that avoid this kind of problem (@pxref{Naming Types,,Naming an
6207 Expression's Type}). However, writing @code{MIN} and @code{MAX} as
6208 macros also forces you to use function-call notation for a
6209 fundamental arithmetic operation. Using GNU C++ extensions, you can
6210 write @w{@samp{int min = i <? j;}} instead.
6212 Since @code{<?} and @code{>?} are built into the compiler, they properly
6213 handle expressions with side-effects; @w{@samp{int min = i++ <? j++;}}
6217 @section When is a Volatile Object Accessed?
6218 @cindex accessing volatiles
6219 @cindex volatile read
6220 @cindex volatile write
6221 @cindex volatile access
6223 Both the C and C++ standard have the concept of volatile objects. These
6224 are normally accessed by pointers and used for accessing hardware. The
6225 standards encourage compilers to refrain from optimizations
6226 concerning accesses to volatile objects that it might perform on
6227 non-volatile objects. The C standard leaves it implementation defined
6228 as to what constitutes a volatile access. The C++ standard omits to
6229 specify this, except to say that C++ should behave in a similar manner
6230 to C with respect to volatiles, where possible. The minimum either
6231 standard specifies is that at a sequence point all previous accesses to
6232 volatile objects have stabilized and no subsequent accesses have
6233 occurred. Thus an implementation is free to reorder and combine
6234 volatile accesses which occur between sequence points, but cannot do so
6235 for accesses across a sequence point. The use of volatiles does not
6236 allow you to violate the restriction on updating objects multiple times
6237 within a sequence point.
6239 In most expressions, it is intuitively obvious what is a read and what is
6240 a write. For instance
6243 volatile int *dst = @var{somevalue};
6244 volatile int *src = @var{someothervalue};
6249 will cause a read of the volatile object pointed to by @var{src} and stores the
6250 value into the volatile object pointed to by @var{dst}. There is no
6251 guarantee that these reads and writes are atomic, especially for objects
6252 larger than @code{int}.
6254 Less obvious expressions are where something which looks like an access
6255 is used in a void context. An example would be,
6258 volatile int *src = @var{somevalue};
6262 With C, such expressions are rvalues, and as rvalues cause a read of
6263 the object, GCC interprets this as a read of the volatile being pointed
6264 to. The C++ standard specifies that such expressions do not undergo
6265 lvalue to rvalue conversion, and that the type of the dereferenced
6266 object may be incomplete. The C++ standard does not specify explicitly
6267 that it is this lvalue to rvalue conversion which is responsible for
6268 causing an access. However, there is reason to believe that it is,
6269 because otherwise certain simple expressions become undefined. However,
6270 because it would surprise most programmers, G++ treats dereferencing a
6271 pointer to volatile object of complete type in a void context as a read
6272 of the object. When the object has incomplete type, G++ issues a
6277 struct T @{int m;@};
6278 volatile S *ptr1 = @var{somevalue};
6279 volatile T *ptr2 = @var{somevalue};
6284 In this example, a warning is issued for @code{*ptr1}, and @code{*ptr2}
6285 causes a read of the object pointed to. If you wish to force an error on
6286 the first case, you must force a conversion to rvalue with, for instance
6287 a static cast, @code{static_cast<S>(*ptr1)}.
6289 When using a reference to volatile, G++ does not treat equivalent
6290 expressions as accesses to volatiles, but instead issues a warning that
6291 no volatile is accessed. The rationale for this is that otherwise it
6292 becomes difficult to determine where volatile access occur, and not
6293 possible to ignore the return value from functions returning volatile
6294 references. Again, if you wish to force a read, cast the reference to
6297 @node Restricted Pointers
6298 @section Restricting Pointer Aliasing
6299 @cindex restricted pointers
6300 @cindex restricted references
6301 @cindex restricted this pointer
6303 As with gcc, g++ understands the C99 feature of restricted pointers,
6304 specified with the @code{__restrict__}, or @code{__restrict} type
6305 qualifier. Because you cannot compile C++ by specifying the @option{-std=c99}
6306 language flag, @code{restrict} is not a keyword in C++.
6308 In addition to allowing restricted pointers, you can specify restricted
6309 references, which indicate that the reference is not aliased in the local
6313 void fn (int *__restrict__ rptr, int &__restrict__ rref)
6320 In the body of @code{fn}, @var{rptr} points to an unaliased integer and
6321 @var{rref} refers to a (different) unaliased integer.
6323 You may also specify whether a member function's @var{this} pointer is
6324 unaliased by using @code{__restrict__} as a member function qualifier.
6327 void T::fn () __restrict__
6334 Within the body of @code{T::fn}, @var{this} will have the effective
6335 definition @code{T *__restrict__ const this}. Notice that the
6336 interpretation of a @code{__restrict__} member function qualifier is
6337 different to that of @code{const} or @code{volatile} qualifier, in that it
6338 is applied to the pointer rather than the object. This is consistent with
6339 other compilers which implement restricted pointers.
6341 As with all outermost parameter qualifiers, @code{__restrict__} is
6342 ignored in function definition matching. This means you only need to
6343 specify @code{__restrict__} in a function definition, rather than
6344 in a function prototype as well.
6347 @section Vague Linkage
6348 @cindex vague linkage
6350 There are several constructs in C++ which require space in the object
6351 file but are not clearly tied to a single translation unit. We say that
6352 these constructs have ``vague linkage''. Typically such constructs are
6353 emitted wherever they are needed, though sometimes we can be more
6357 @item Inline Functions
6358 Inline functions are typically defined in a header file which can be
6359 included in many different compilations. Hopefully they can usually be
6360 inlined, but sometimes an out-of-line copy is necessary, if the address
6361 of the function is taken or if inlining fails. In general, we emit an
6362 out-of-line copy in all translation units where one is needed. As an
6363 exception, we only emit inline virtual functions with the vtable, since
6364 it will always require a copy.
6366 Local static variables and string constants used in an inline function
6367 are also considered to have vague linkage, since they must be shared
6368 between all inlined and out-of-line instances of the function.
6372 C++ virtual functions are implemented in most compilers using a lookup
6373 table, known as a vtable. The vtable contains pointers to the virtual
6374 functions provided by a class, and each object of the class contains a
6375 pointer to its vtable (or vtables, in some multiple-inheritance
6376 situations). If the class declares any non-inline, non-pure virtual
6377 functions, the first one is chosen as the ``key method'' for the class,
6378 and the vtable is only emitted in the translation unit where the key
6381 @emph{Note:} If the chosen key method is later defined as inline, the
6382 vtable will still be emitted in every translation unit which defines it.
6383 Make sure that any inline virtuals are declared inline in the class
6384 body, even if they are not defined there.
6386 @item type_info objects
6389 C++ requires information about types to be written out in order to
6390 implement @samp{dynamic_cast}, @samp{typeid} and exception handling.
6391 For polymorphic classes (classes with virtual functions), the type_info
6392 object is written out along with the vtable so that @samp{dynamic_cast}
6393 can determine the dynamic type of a class object at runtime. For all
6394 other types, we write out the type_info object when it is used: when
6395 applying @samp{typeid} to an expression, throwing an object, or
6396 referring to a type in a catch clause or exception specification.
6398 @item Template Instantiations
6399 Most everything in this section also applies to template instantiations,
6400 but there are other options as well.
6401 @xref{Template Instantiation,,Where's the Template?}.
6405 When used with GNU ld version 2.8 or later on an ELF system such as
6406 Linux/GNU or Solaris 2, or on Microsoft Windows, duplicate copies of
6407 these constructs will be discarded at link time. This is known as
6410 On targets that don't support COMDAT, but do support weak symbols, GCC
6411 will use them. This way one copy will override all the others, but
6412 the unused copies will still take up space in the executable.
6414 For targets which do not support either COMDAT or weak symbols,
6415 most entities with vague linkage will be emitted as local symbols to
6416 avoid duplicate definition errors from the linker. This will not happen
6417 for local statics in inlines, however, as having multiple copies will
6418 almost certainly break things.
6420 @xref{C++ Interface,,Declarations and Definitions in One Header}, for
6421 another way to control placement of these constructs.
6424 @section Declarations and Definitions in One Header
6426 @cindex interface and implementation headers, C++
6427 @cindex C++ interface and implementation headers
6428 C++ object definitions can be quite complex. In principle, your source
6429 code will need two kinds of things for each object that you use across
6430 more than one source file. First, you need an @dfn{interface}
6431 specification, describing its structure with type declarations and
6432 function prototypes. Second, you need the @dfn{implementation} itself.
6433 It can be tedious to maintain a separate interface description in a
6434 header file, in parallel to the actual implementation. It is also
6435 dangerous, since separate interface and implementation definitions may
6436 not remain parallel.
6438 @cindex pragmas, interface and implementation
6439 With GNU C++, you can use a single header file for both purposes.
6442 @emph{Warning:} The mechanism to specify this is in transition. For the
6443 nonce, you must use one of two @code{#pragma} commands; in a future
6444 release of GNU C++, an alternative mechanism will make these
6445 @code{#pragma} commands unnecessary.
6448 The header file contains the full definitions, but is marked with
6449 @samp{#pragma interface} in the source code. This allows the compiler
6450 to use the header file only as an interface specification when ordinary
6451 source files incorporate it with @code{#include}. In the single source
6452 file where the full implementation belongs, you can use either a naming
6453 convention or @samp{#pragma implementation} to indicate this alternate
6454 use of the header file.
6457 @item #pragma interface
6458 @itemx #pragma interface "@var{subdir}/@var{objects}.h"
6459 @kindex #pragma interface
6460 Use this directive in @emph{header files} that define object classes, to save
6461 space in most of the object files that use those classes. Normally,
6462 local copies of certain information (backup copies of inline member
6463 functions, debugging information, and the internal tables that implement
6464 virtual functions) must be kept in each object file that includes class
6465 definitions. You can use this pragma to avoid such duplication. When a
6466 header file containing @samp{#pragma interface} is included in a
6467 compilation, this auxiliary information will not be generated (unless
6468 the main input source file itself uses @samp{#pragma implementation}).
6469 Instead, the object files will contain references to be resolved at link
6472 The second form of this directive is useful for the case where you have
6473 multiple headers with the same name in different directories. If you
6474 use this form, you must specify the same string to @samp{#pragma
6477 @item #pragma implementation
6478 @itemx #pragma implementation "@var{objects}.h"
6479 @kindex #pragma implementation
6480 Use this pragma in a @emph{main input file}, when you want full output from
6481 included header files to be generated (and made globally visible). The
6482 included header file, in turn, should use @samp{#pragma interface}.
6483 Backup copies of inline member functions, debugging information, and the
6484 internal tables used to implement virtual functions are all generated in
6485 implementation files.
6487 @cindex implied @code{#pragma implementation}
6488 @cindex @code{#pragma implementation}, implied
6489 @cindex naming convention, implementation headers
6490 If you use @samp{#pragma implementation} with no argument, it applies to
6491 an include file with the same basename@footnote{A file's @dfn{basename}
6492 was the name stripped of all leading path information and of trailing
6493 suffixes, such as @samp{.h} or @samp{.C} or @samp{.cc}.} as your source
6494 file. For example, in @file{allclass.cc}, giving just
6495 @samp{#pragma implementation}
6496 by itself is equivalent to @samp{#pragma implementation "allclass.h"}.
6498 In versions of GNU C++ prior to 2.6.0 @file{allclass.h} was treated as
6499 an implementation file whenever you would include it from
6500 @file{allclass.cc} even if you never specified @samp{#pragma
6501 implementation}. This was deemed to be more trouble than it was worth,
6502 however, and disabled.
6504 If you use an explicit @samp{#pragma implementation}, it must appear in
6505 your source file @emph{before} you include the affected header files.
6507 Use the string argument if you want a single implementation file to
6508 include code from multiple header files. (You must also use
6509 @samp{#include} to include the header file; @samp{#pragma
6510 implementation} only specifies how to use the file---it doesn't actually
6513 There is no way to split up the contents of a single header file into
6514 multiple implementation files.
6517 @cindex inlining and C++ pragmas
6518 @cindex C++ pragmas, effect on inlining
6519 @cindex pragmas in C++, effect on inlining
6520 @samp{#pragma implementation} and @samp{#pragma interface} also have an
6521 effect on function inlining.
6523 If you define a class in a header file marked with @samp{#pragma
6524 interface}, the effect on a function defined in that class is similar to
6525 an explicit @code{extern} declaration---the compiler emits no code at
6526 all to define an independent version of the function. Its definition
6527 is used only for inlining with its callers.
6529 @opindex fno-implement-inlines
6530 Conversely, when you include the same header file in a main source file
6531 that declares it as @samp{#pragma implementation}, the compiler emits
6532 code for the function itself; this defines a version of the function
6533 that can be found via pointers (or by callers compiled without
6534 inlining). If all calls to the function can be inlined, you can avoid
6535 emitting the function by compiling with @option{-fno-implement-inlines}.
6536 If any calls were not inlined, you will get linker errors.
6538 @node Template Instantiation
6539 @section Where's the Template?
6541 @cindex template instantiation
6543 C++ templates are the first language feature to require more
6544 intelligence from the environment than one usually finds on a UNIX
6545 system. Somehow the compiler and linker have to make sure that each
6546 template instance occurs exactly once in the executable if it is needed,
6547 and not at all otherwise. There are two basic approaches to this
6548 problem, which I will refer to as the Borland model and the Cfront model.
6552 Borland C++ solved the template instantiation problem by adding the code
6553 equivalent of common blocks to their linker; the compiler emits template
6554 instances in each translation unit that uses them, and the linker
6555 collapses them together. The advantage of this model is that the linker
6556 only has to consider the object files themselves; there is no external
6557 complexity to worry about. This disadvantage is that compilation time
6558 is increased because the template code is being compiled repeatedly.
6559 Code written for this model tends to include definitions of all
6560 templates in the header file, since they must be seen to be
6564 The AT&T C++ translator, Cfront, solved the template instantiation
6565 problem by creating the notion of a template repository, an
6566 automatically maintained place where template instances are stored. A
6567 more modern version of the repository works as follows: As individual
6568 object files are built, the compiler places any template definitions and
6569 instantiations encountered in the repository. At link time, the link
6570 wrapper adds in the objects in the repository and compiles any needed
6571 instances that were not previously emitted. The advantages of this
6572 model are more optimal compilation speed and the ability to use the
6573 system linker; to implement the Borland model a compiler vendor also
6574 needs to replace the linker. The disadvantages are vastly increased
6575 complexity, and thus potential for error; for some code this can be
6576 just as transparent, but in practice it can been very difficult to build
6577 multiple programs in one directory and one program in multiple
6578 directories. Code written for this model tends to separate definitions
6579 of non-inline member templates into a separate file, which should be
6580 compiled separately.
6583 When used with GNU ld version 2.8 or later on an ELF system such as
6584 Linux/GNU or Solaris 2, or on Microsoft Windows, g++ supports the
6585 Borland model. On other systems, g++ implements neither automatic
6588 A future version of g++ will support a hybrid model whereby the compiler
6589 will emit any instantiations for which the template definition is
6590 included in the compile, and store template definitions and
6591 instantiation context information into the object file for the rest.
6592 The link wrapper will extract that information as necessary and invoke
6593 the compiler to produce the remaining instantiations. The linker will
6594 then combine duplicate instantiations.
6596 In the mean time, you have the following options for dealing with
6597 template instantiations:
6602 Compile your template-using code with @option{-frepo}. The compiler will
6603 generate files with the extension @samp{.rpo} listing all of the
6604 template instantiations used in the corresponding object files which
6605 could be instantiated there; the link wrapper, @samp{collect2}, will
6606 then update the @samp{.rpo} files to tell the compiler where to place
6607 those instantiations and rebuild any affected object files. The
6608 link-time overhead is negligible after the first pass, as the compiler
6609 will continue to place the instantiations in the same files.
6611 This is your best option for application code written for the Borland
6612 model, as it will just work. Code written for the Cfront model will
6613 need to be modified so that the template definitions are available at
6614 one or more points of instantiation; usually this is as simple as adding
6615 @code{#include <tmethods.cc>} to the end of each template header.
6617 For library code, if you want the library to provide all of the template
6618 instantiations it needs, just try to link all of its object files
6619 together; the link will fail, but cause the instantiations to be
6620 generated as a side effect. Be warned, however, that this may cause
6621 conflicts if multiple libraries try to provide the same instantiations.
6622 For greater control, use explicit instantiation as described in the next
6626 @opindex fno-implicit-templates
6627 Compile your code with @option{-fno-implicit-templates} to disable the
6628 implicit generation of template instances, and explicitly instantiate
6629 all the ones you use. This approach requires more knowledge of exactly
6630 which instances you need than do the others, but it's less
6631 mysterious and allows greater control. You can scatter the explicit
6632 instantiations throughout your program, perhaps putting them in the
6633 translation units where the instances are used or the translation units
6634 that define the templates themselves; you can put all of the explicit
6635 instantiations you need into one big file; or you can create small files
6642 template class Foo<int>;
6643 template ostream& operator <<
6644 (ostream&, const Foo<int>&);
6647 for each of the instances you need, and create a template instantiation
6650 If you are using Cfront-model code, you can probably get away with not
6651 using @option{-fno-implicit-templates} when compiling files that don't
6652 @samp{#include} the member template definitions.
6654 If you use one big file to do the instantiations, you may want to
6655 compile it without @option{-fno-implicit-templates} so you get all of the
6656 instances required by your explicit instantiations (but not by any
6657 other files) without having to specify them as well.
6659 g++ has extended the template instantiation syntax outlined in the
6660 Working Paper to allow forward declaration of explicit instantiations
6661 (with @code{extern}), instantiation of the compiler support data for a
6662 template class (i.e.@: the vtable) without instantiating any of its
6663 members (with @code{inline}), and instantiation of only the static data
6664 members of a template class, without the support data or member
6665 functions (with (@code{static}):
6668 extern template int max (int, int);
6669 inline template class Foo<int>;
6670 static template class Foo<int>;
6674 Do nothing. Pretend g++ does implement automatic instantiation
6675 management. Code written for the Borland model will work fine, but
6676 each translation unit will contain instances of each of the templates it
6677 uses. In a large program, this can lead to an unacceptable amount of code
6681 @opindex fexternal-templates
6682 Add @samp{#pragma interface} to all files containing template
6683 definitions. For each of these files, add @samp{#pragma implementation
6684 "@var{filename}"} to the top of some @samp{.C} file which
6685 @samp{#include}s it. Then compile everything with
6686 @option{-fexternal-templates}. The templates will then only be expanded
6687 in the translation unit which implements them (i.e.@: has a @samp{#pragma
6688 implementation} line for the file where they live); all other files will
6689 use external references. If you're lucky, everything should work
6690 properly. If you get undefined symbol errors, you need to make sure
6691 that each template instance which is used in the program is used in the
6692 file which implements that template. If you don't have any use for a
6693 particular instance in that file, you can just instantiate it
6694 explicitly, using the syntax from the latest C++ working paper:
6697 template class A<int>;
6698 template ostream& operator << (ostream&, const A<int>&);
6701 This strategy will work with code written for either model. If you are
6702 using code written for the Cfront model, the file containing a class
6703 template and the file containing its member templates should be
6704 implemented in the same translation unit.
6707 @opindex falt-external-templates
6708 A slight variation on this approach is to use the flag
6709 @option{-falt-external-templates} instead. This flag causes template
6710 instances to be emitted in the translation unit that implements the
6711 header where they are first instantiated, rather than the one which
6712 implements the file where the templates are defined. This header must
6713 be the same in all translation units, or things are likely to break.
6715 @xref{C++ Interface,,Declarations and Definitions in One Header}, for
6716 more discussion of these pragmas.
6719 @node Bound member functions
6720 @section Extracting the function pointer from a bound pointer to member function
6723 @cindex pointer to member function
6724 @cindex bound pointer to member function
6726 In C++, pointer to member functions (PMFs) are implemented using a wide
6727 pointer of sorts to handle all the possible call mechanisms; the PMF
6728 needs to store information about how to adjust the @samp{this} pointer,
6729 and if the function pointed to is virtual, where to find the vtable, and
6730 where in the vtable to look for the member function. If you are using
6731 PMFs in an inner loop, you should really reconsider that decision. If
6732 that is not an option, you can extract the pointer to the function that
6733 would be called for a given object/PMF pair and call it directly inside
6734 the inner loop, to save a bit of time.
6736 Note that you will still be paying the penalty for the call through a
6737 function pointer; on most modern architectures, such a call defeats the
6738 branch prediction features of the CPU@. This is also true of normal
6739 virtual function calls.
6741 The syntax for this extension is
6745 extern int (A::*fp)();
6746 typedef int (*fptr)(A *);
6748 fptr p = (fptr)(a.*fp);
6751 For PMF constants (i.e.@: expressions of the form @samp{&Klasse::Member}),
6752 no object is needed to obtain the address of the function. They can be
6753 converted to function pointers directly:
6756 fptr p1 = (fptr)(&A::foo);
6759 @opindex Wno-pmf-conversions
6760 You must specify @option{-Wno-pmf-conversions} to use this extension.
6762 @node C++ Attributes
6763 @section C++-Specific Variable, Function, and Type Attributes
6765 Some attributes only make sense for C++ programs.
6768 @item init_priority (@var{priority})
6769 @cindex init_priority attribute
6772 In Standard C++, objects defined at namespace scope are guaranteed to be
6773 initialized in an order in strict accordance with that of their definitions
6774 @emph{in a given translation unit}. No guarantee is made for initializations
6775 across translation units. However, GNU C++ allows users to control the
6776 order of initialization of objects defined at namespace scope with the
6777 @code{init_priority} attribute by specifying a relative @var{priority},
6778 a constant integral expression currently bounded between 101 and 65535
6779 inclusive. Lower numbers indicate a higher priority.
6781 In the following example, @code{A} would normally be created before
6782 @code{B}, but the @code{init_priority} attribute has reversed that order:
6785 Some_Class A __attribute__ ((init_priority (2000)));
6786 Some_Class B __attribute__ ((init_priority (543)));
6790 Note that the particular values of @var{priority} do not matter; only their
6793 @item java_interface
6794 @cindex java_interface attribute
6796 This type attribute informs C++ that the class is a Java interface. It may
6797 only be applied to classes declared within an @code{extern "Java"} block.
6798 Calls to methods declared in this interface will be dispatched using GCJ's
6799 interface table mechanism, instead of regular virtual table dispatch.
6803 @node Java Exceptions
6804 @section Java Exceptions
6806 The Java language uses a slightly different exception handling model
6807 from C++. Normally, GNU C++ will automatically detect when you are
6808 writing C++ code that uses Java exceptions, and handle them
6809 appropriately. However, if C++ code only needs to execute destructors
6810 when Java exceptions are thrown through it, GCC will guess incorrectly.
6811 Sample problematic code is:
6814 struct S @{ ~S(); @};
6815 extern void bar(); // is written in Java, and may throw exceptions
6824 The usual effect of an incorrect guess is a link failure, complaining of
6825 a missing routine called @samp{__gxx_personality_v0}.
6827 You can inform the compiler that Java exceptions are to be used in a
6828 translation unit, irrespective of what it might think, by writing
6829 @samp{@w{#pragma GCC java_exceptions}} at the head of the file. This
6830 @samp{#pragma} must appear before any functions that throw or catch
6831 exceptions, or run destructors when exceptions are thrown through them.
6833 You cannot mix Java and C++ exceptions in the same translation unit. It
6834 is believed to be safe to throw a C++ exception from one file through
6835 another file compiled for the Java exception model, or vice versa, but
6836 there may be bugs in this area.
6838 @node Deprecated Features
6839 @section Deprecated Features
6841 In the past, the GNU C++ compiler was extended to experiment with new
6842 features, at a time when the C++ language was still evolving. Now that
6843 the C++ standard is complete, some of those features are superseded by
6844 superior alternatives. Using the old features might cause a warning in
6845 some cases that the feature will be dropped in the future. In other
6846 cases, the feature might be gone already.
6848 While the list below is not exhaustive, it documents some of the options
6849 that are now deprecated:
6852 @item -fexternal-templates
6853 @itemx -falt-external-templates
6854 These are two of the many ways for g++ to implement template
6855 instantiation. @xref{Template Instantiation}. The C++ standard clearly
6856 defines how template definitions have to be organized across
6857 implementation units. g++ has an implicit instantiation mechanism that
6858 should work just fine for standard-conforming code.
6860 @item -fstrict-prototype
6861 @itemx -fno-strict-prototype
6862 Previously it was possible to use an empty prototype parameter list to
6863 indicate an unspecified number of parameters (like C), rather than no
6864 parameters, as C++ demands. This feature has been removed, except where
6865 it is required for backwards compatibility @xref{Backwards Compatibility}.
6868 The named return value extension has been deprecated, and is now
6871 The use of initializer lists with new expressions has been deprecated,
6872 and is now removed from g++.
6874 Floating and complex non-type template parameters have been deprecated,
6875 and are now removed from g++.
6877 The implicit typename extension has been deprecated and will be removed
6878 from g++ at some point. In some cases g++ determines that a dependant
6879 type such as @code{TPL<T>::X} is a type without needing a
6880 @code{typename} keyword, contrary to the standard.
6882 @node Backwards Compatibility
6883 @section Backwards Compatibility
6884 @cindex Backwards Compatibility
6885 @cindex ARM [Annotated C++ Reference Manual]
6887 Now that there is a definitive ISO standard C++, G++ has a specification
6888 to adhere to. The C++ language evolved over time, and features that
6889 used to be acceptable in previous drafts of the standard, such as the ARM
6890 [Annotated C++ Reference Manual], are no longer accepted. In order to allow
6891 compilation of C++ written to such drafts, G++ contains some backwards
6892 compatibilities. @emph{All such backwards compatibility features are
6893 liable to disappear in future versions of G++.} They should be considered
6894 deprecated @xref{Deprecated Features}.
6898 If a variable is declared at for scope, it used to remain in scope until
6899 the end of the scope which contained the for statement (rather than just
6900 within the for scope). G++ retains this, but issues a warning, if such a
6901 variable is accessed outside the for scope.
6903 @item Implicit C language
6904 Old C system header files did not contain an @code{extern "C" @{@dots{}@}}
6905 scope to set the language. On such systems, all header files are
6906 implicitly scoped inside a C language scope. Also, an empty prototype
6907 @code{()} will be treated as an unspecified number of arguments, rather
6908 than no arguments, as C++ demands.